Preparative Isolation of High-Purity n-3 Docosapentaenoic Acid via Iterative Isocratic Flash Chromatography with Solvent Recycling

Abstract

Background: n-3 Docosapentaenoic acid (DPA; 22:5 n-3) is increasingly viewed as a distinct long-chain omega-3 fatty acid with biological activities that are not fully captured by eicosapentaenoic acid (EPA) or docosahexaenoic acid (DHA). However, progress remains limited by restricted access to high-purity DPA: most commercial sources contain DPA as a minor component, and published isolation strategies often yield only enriched mixtures or require multi-step workflows that are difficult to scale in standard laboratories. Objectives: We aimed to establish a robust, laboratory-accessible purification workflow to obtain DPA ethyl ester at high purity while preserving oxidative quality. Methods: Candidate lipid sources were screened to select an optimal DPA-containing feedstock. Oils were stabilized with antioxidants and pre-fractionated by cold crystallization (−20 °C) to reduce saturated lipids and oxidation by-products. Preparative separation used a stacked C18 flash system (15 μm + 45 μm in series) operated isocratically (methanol/water 92:8, v/v) at 120 mL/min. Fractions were analyzed by GC and iteratively reinjected to progressively enrich the DPA window. Solvent was recovered by distillation and reused. Results: Omegavie^® 4020EE (5.4% n-3 DPA) was identified as the best starting material. Pretreatment eliminated detectable TBARS-derived malondialdehyde. The isocratic purification-loop strategy produced tens of grams of DPA ethyl ester at >98% purity (GC–FID) defined as n-3 DPA area% of total identified fatty acid methyl esters by GC–FID, with per-cycle DPA recovery of 91–95%, overall recovery of 76% from the starting DPA content, and >90% solvent recycling. The workflow is scalable at the gram-to-tens-of-grams level for research laboratories, although solvent burden and column maintenance remain practical constraints for larger-scale implementation. Identity and purity were confirmed by GC–MS and ^1H NMR, and oxidation indices remained low (peroxide value < 0.2 meq/kg; p-anisidine < 3). Conclusions: This scalable, solvent-conscious protocol enables reliable access to high-purity DPA and should be adaptable to other low-abundance polyunsaturated fatty acids.

1. Introduction

The long-chain omega-3 polyunsaturated fatty acids eicosapentaenoic acid (EPA, 20:5 n-3) and docosahexaenoic acid (DHA, 22:6 n-3) are well-known for their biological effects, whereas n-3 docosapentaenoic acid (DPA, 22:5 n-3) has been comparatively overlooked. Recent reviews highlight DPA as the “missing” omega-3 fatty acid, noting its presence in marine foods and mammalian tissues. DPA bridges the metabolic pathway between EPA and DHA and may exert its own effects, for instance, via conversion to specialized pro-resolving mediators distinct from those derived from EPA or DHA. Emerging evidence suggests DPA has anti-inflammatory and cardiometabolic benefits, reinforcing its interest [1,2,3,4].

DPA occurs in fish oils, marine animal oils, human milk, and some terrestrial tissues, but usually at low abundance relative to EPA and DHA. In most natural or commercial lipid matrices, it remains a minor component, which has limited routine access to purified DPA for mechanistic, analytical, and biological studies [1,2,4]. Past efforts to purify DPA have been scarce and have generally involved multi-step protocols. Notably, Yamamura and Shimomura achieved >99% pure DPA (and DHA) from single-cell oil using industrial-scale HPLC columns in series, with a reported production rate of ~70 g DPA per hour [5]. This impressive output required specialized equipment, twin 400 mm i.d. columns and high-pressure pumps, not commonly accessible for academic labs. At the laboratory scale, Mu et al. employed a combination of urea complexation and argentated silica column chromatography to enrich DPA from tuna oil, obtaining a fraction with ~22% DPA and an overall DPA yield of ~70% [6]. While this approach confirmed the feasibility of concentrating DPA, it fell short of producing gram-quantities of pure DPA. More recently, other groups have reported partial enrichment of DPA by low-temperature crystallization or as specific derivatives—for instance, concentrating DPA as a diacylglycerol from microbial oils—but still did not reach high purity or scalability [7,8,9]. A few patents have likewise been filed for DPA production or isolation methods (e.g., via algal sources or modified oils), underscoring the technical interest in this problem [5,6,7,8,9,10,11].

In summary, there is a clear need for a practical, high-purity DPA isolation method that can be implemented with standard laboratory equipment. That need is not only practical but also physicochemical. In preparative reversed-phase systems, DPA must be separated from structurally similar long-chain polyunsaturated fatty acids that differ only subtly in chain length and unsaturation. Under overloaded conditions, these small differences are insufficient to guarantee baseline resolution, especially for neighboring species such as DHA and n-6 DPA. Moreover, the high unsaturation of DPA increases susceptibility to oxidation and may favor concentration-dependent association phenomena that broaden bands and complicate repeated enrichment cycles. Accordingly, solvent-strength matching, controlled loading, and iterative selection of the central DPA-rich window are not merely operational choices but mechanistically important features of the method. Here, we present a preparative chromatography protocol designed to yield tens of grams of DPA (ethyl ester form) at >98% purity. The method uses readily available C18 reversed-phase columns in a stacked configuration and an iterative purification loop strategy to incrementally enrich DPA. Importantly, the process incorporates solvent recycling and oxidative precautions to enhance sustainability and product quality. We detail the optimization of key parameters such as feedstock selection, flow rate, isocratic elution conditions, loading capacity, and fractionation strategy. We also compare the performance to prior ω-3 PUFA purification approaches. The specific novelty of this work is not a new stationary phase or separation theory, but the demonstration that high-purity n-3 DPA can be obtained using standard laboratory C18 flash hardware through an iterative, windowed enrichment strategy that combines overloaded isocratic separations, solvent-strength control during reinjection, oxidation management, and solvent recycling. This work provides an enabling technique to supply high-purity DPA for nutritional biochemistry, lipid mediator research, and other applications that have been hampered by DPA’s limited availability.

2. Materials and Methods

2.1. Materials

Commercial long-chain omega-3 fatty acid concentrates were obtained from Polaris (Quimper, France). The primary material was a fish oil ethyl ester blend (Omegavie 4020EE, Polaris, Quimper, France) containing approximately 40% total omega-3, enriched in EPA, DHA, and DPA. For source-screening and method-development comparisons, additional commercial concentrates (Omegavie DHA80, Omegavie DHA70, and 1812 TG, all from Polaris, Quimper, France) and tuna oil were also tested. All oils were stored in amber containers under nitrogen. Butylated hydroxytoluene (BHT, 0.02% w/v) was added as an antioxidant. Beef liver was examined only in preliminary exploratory extraction trials and was not included in the comparative source panel reported in

Table 1. Composition of food sources used for DPA purification (wt% of fatty acids).

Fatty Acid	DHA 80	DHA 70	1812 TG	4020 EE	Tuna
Saturated	0	6.2	27.5	4.9	27.5
Monounsaturated	0.5	8.2	24.4	12.3	23.3
C16:2 n-4	0	0	1.1	1.1	0.3
C16:3 n-4	0	0	1.2	0	0.2
C16:4 n-3	0	0	0.1	0	0.1
C16:4 isomer	0	0	2.2	0	0.2
C18:2 n-6	0	0.4	0.4	0.4	2
C18:3 n-6	0	0	0	0.2	0.4
C18:3 n-3	0.1	0	0.1	0.1	0.6
C18:4 n-3	0.6	0.9	0.2	0.5	0.9
C20:2 n-6	0.1	0	0.2	0.1	0.3
C20:3 n-9	0	0	0	0	0.1
C20:3 n-6	0	0	0.1	0	0.2
C20:4 n-6 (ARA)	0.1	0.7	0.9	1.5	2.2
C20:3 n-3	0	0	0	0.1	0.2
C20:4 n-3	0.8	0.1	0.2	0.1	0.2
C20:5 n-3 (EPA)	2.8	4.1	19.7	45	8.5
C21:5 n-3	0.2	0.2	0.8	0	0.2
C22:5 n-6 (DPA)	0.5	0.7	0.3	0.5	1.4
C22:5 n-3 (DPA)	5.2	4.1	0.3	5.4	1.5
C22:6 n-3 (DHA)	84.4	69	12.8	23.8	29.7

. HPLC-grade methanol and analytical-grade solvents were purchased from Fisher Scientific (Pittsburgh, PA, USA), and Milli-Q (Merck, Rahway, NJ, USA) deionized water was used throughout. Boron trifluoride methanol reagent (14% BF₃) for methylation was from Sigma-Aldrich (Gillingham, UK).

Fatty Acid Analysis

The fatty acid profiles of oils and tissue extracts were determined by gas chromatography–mass spectrometry (GC–MS) of fatty acid methyl esters (FAME). Lipids from oils or homogenized tissues were extracted according to Folch et al. with minor modifications. Briefly, samples were mixed with chloroform/methanol (2:1, v/v), homogenized thoroughly, and washed with aqueous phase to induce partitioning. The lower organic phase was collected and evaporated under nitrogen, and the recovered lipids were then subjected to saponification and methylation. Saponification (0.5 N NaOH in methanol, 70 °C, 30 min) was followed by methylation using BF₃–methanol (70 °C, 30 min) [12]. The resulting FAME was extracted in hexane and analyzed on an Agilent 7890A GC coupled to a 5975C MS detector (Agilent, Santa Clara, CA, USA). A polar capillary column (BPX70, 60 m × 0.25 mm, 0.25 μm film) was used with helium as a carrier gas (1.8 mL/min). The oven temperature was programmed from 150 °C, ramping 4 °C/min to 250 °C (hold 10 min). MS was operated in EI mode (70 eV) scanning m/z 50–550. Individual fatty acids were identified by retention time matching to authentic standards and by mass spectral library (NIST) matching for unknowns. Quantification was carried out by FID (flame ionization detection) on a parallel GC system using a 100 m CP-Sil 88 column, with C17:0 internal standard for calibration. All fatty acid compositions are reported as weight percent of total identified FAME.

2.2. Chromatography Setup

Preparative purification was performed with a modular flash chromatography system (Interchim puriFlash^® (Interchim, Montluçon, France) equipped with a binary pump and UV detector) modified to accommodate high flow rates. Two C18 reversed-phase columns were connected in series: Column 1 (high-capacity) was a 45 μm silica C18 cartridge (approx. 250 × 100 mm, ~330 g packing), and Column 2 (high-performance) was a 15 μm C18 cartridge of the same dimensions. The combined bed length thus provided enhanced resolution and load tolerance. The mobile phase was methanol–water with 0.1% formic acid (to aid in UV detection and maintain consistency, though formic acid is not strictly required). Initial scouting runs compared 95:5 vs. 92:8 MeOH/H₂O; the latter was adopted for production runs as described. The flow rate was set to 120 mL/min, and the system pressure stabilized around 200 bar (well within column limits). The temperature was ambient (20–25 °C).

An injection loop of 20 mL volume was used to introduce the sample, which consisted of the fish oil ethyl ester concentrate Omegavie 4020EE^® diluted 1:10 (v/v) in methanol. For the initial crude-feed injections, this strong methanolic feed was accepted because the early cycles were designed primarily for bulk EPA removal rather than maximal DHA–DPA resolution. For reinjection cycles, pooled DPA-rich fractions were solvent-matched to the starting mobile-phase composition before loading to minimize strong solvent distortion. Each injection thus contained ~1.5–2.0 g of oil (depending on oil density and composition). This value refers to the mass of 4020EE oil present within the 20 mL injected solution, not to 20 mL of neat oil.

Fractions were collected into glass tubes using an automated fraction collector triggered by time. We collected 60 mL per fraction (approximately every 30 s) based on prior calibration of the system dead volume and peak elution profile. The UV detector was set at 205 nm to monitor elution; distinct increases in absorbance signaled the EPA, DPA, and DHA band passages, but GC analysis was used for definitive fraction mapping.

2.3. Iterative Purification Loop

After each 20 mL injection, collected fractions were processed as follows. Methanol was evaporated from each tube under vacuum or nitrogen sweep, and the residue was weighed to determine recovery. A small aliquot of each dried fraction was derivatized to FAME and analyzed by GC–FID to determine its DPA content. Fractions were then grouped into four pools corresponding to the DHA-rich region (pool A), the transitional region before DPA (B), the DPA-rich peak (C), and the trailing region including n-6 DPA (D). Only pool C fractions (typically 4–6 tubes per run) were retained for the next purification cycle. Prior to reinjection, the pooled C fractions were combined and redissolved in methanol (up to 20 mL total) and an equal volume of water added to match the 92:8 mobile phase to avoid a strong solvent effect on the column. This solution was then injected as the feed for the next cycle. The columns were not regenerated between cycles beyond a brief flush with fresh mobile phase (15 min) to re-equilibrate isocratically. This cycle was repeated until the GC–FID analysis of pool C showed no detectable DHA. In practice, three iterations were sufficient to reach >98% purity for DPA. After the final cycle, the purified DPA fractions were combined and stored under nitrogen at −20 °C until use.

For maximum recovery, the pool D fractions containing DPA and n-6 DPA isomer can be further processed. In our case, we performed an additional polishing step on these fractions by using a preparative silver-ion column (25 cm × 2.0 cm i.d. glass column packed with 5% AgNO₃-silica). A 1 mL portion of the pool D concentrate was loaded and eluted with 95:5 methanol/water at 120 mL/min, which cleanly separated n-3 DPA from n-6 DPA (confirmed by GC). The collected fraction from this optional polishing step contained n-3 DPA at >99 area% of total identified FAME by GC–FID and was then combined with the main batch. In this context, the >99% value refers to the resolved n-3 DPA fraction obtained after separation from the terminal n-6 DPA-containing pool, not to the unresolved combined DPA isomer mixture. This optional step, however, added time and was only used to maximize final purity for analytical standard preparation. Most of the results reported were obtained without this silver column step, i.e., with final purity ~98% and a minor n-6 DPA presence (<1–2%).

2.4. Solvent Recycling

To curtail solvent use, we implemented a distillation recovery system. The collected methanol–water effluent from the flash system was funneled into a large round-bottom flask. Using a rotary evaporator, we distilled the bulk methanol (boiling ~65 °C) from the water. Because the mobile phase was only 8% water, once 92% of the volume was collected as distillate, the remainder was water-rich and was discarded. The distilled methanol was dried over anhydrous sodium sulfate and reused for subsequent mobile phase makeup, after verifying that it produced no change in chromatographic behavior. All glassware and tubing in contact with solvents or lipids were rinsed with ethanol and acetone and dried between runs to prevent cross-contamination or microbial growth. Waste methanol from final residues was collected and handled according to institutional safety protocols [13,14].

2.5. Analytical and Quality Control

The purity of the final DPA preparation was assessed by multiple analytical techniques. Quantitative GC–FID was the primary measure of purity, reported as the area percentage of n-3 DPA relative to total identified fatty acid methyl esters (FAME) in the analyzed sample. GC–MS was used qualitatively to check for any unidentified components co-eluting with DPA. ^1H NMR spectra of the purified DPA ethyl ester were recorded on a 400 MHz spectrometer in CDCl₃ to confirm structural identity. Key NMR signals (triplets from terminal methyl protons, bis-allylic methylene protons around 2.8 ppm, and other methylene envelope patterns) matched the literature values for 22:5 n-3. Oxidation metrics were measured on both the starting material and final product: peroxide value (PV) by iodometric titration (AOCS Official Method Cd 8b-90) [15], p-anisidine value by the AOCS Cd 18–90 colorimetric method [16], and TBARS by the standard thiobarbituric acid assay (using malondialdehyde equivalents). In addition, an HPLC-fluorescence method for MDA (after derivatization to a fluorescent adduct) was used on select samples to verify TBARS results, following established protocols [15]. UV absorbance spectra (200–300 nm) of the oils were recorded to detect conjugated diene or triene formation (signs of polyunsaturated lipid oxidation). All analyses confirmed that the processing did not introduce oxidation: PV remained <0.5 meq/kg, and no significant increase in UV absorbance at 233 nm or 270 nm was observed in the purified DPA compared to the fresh oil.

2.6. Statistical Note

Purity, recovery, and oxidation metrics are reported as mean ± standard deviation from at least three independent purification runs unless otherwise indicated. Screening data used only for feedstock selection are presented descriptively because biological replication was not applicable. Operating-condition comparisons are therefore discussed descriptively unless formal replicated statistical testing is explicitly reported.

3. Results

3.1. Preparation and Selection of Food Sources

Although oils were stored under nitrogen, we detected trace lipid peroxidation (malondialdehyde, TBARS), consistent with the oxidative liability of highly unsaturated ethyl esters. A single −20 °C crystallization step removed aggregated residues including saturated lipids and oxidation by-products, after which TBARS-derived malondialdehyde was no longer detectable in the supernatant. We next profiled candidate sources (fish oils and enriched n-3 LCPUFA concentrates) to identify an input that maximized DPA while limiting species known to compromise reversed-phase resolution. Across the screened panel reported in Table 1, n-3 DPA ranged from 0.3% (1812 TG) to 5.4% (Omegavie 4020EE^®, Polaris) of total fatty acids (Table 1).

Table 1 reports descriptive screening compositions used for feedstock selection; inferential statistics were not applied because the table summarizes analytical characterization rather than replicated treatment groups.

We selected 4020EE because it combined high DPA with lower DHA and n-6 DPA (22:5n-6). The two principal interferents partially co-elute with DPA on C18 while containing abundant EPA, which elutes well ahead of DPA and can be stripped early (Figure 1).

Conversely, DHA-enriched oil (DHA80^®) was deliberately used to stress-test the separation window, because under overloaded conditions the broad DHA band encroaches into the DPA region and defines the practical worst case for preparative resolution (Figure 1).

3.2. Selection of Columns, Scale-Up, and Elution Optimization

On the small scale (15 µm C18 flash format; MeOH/H₂O 95/5), the HP column gave the best DHA–DPA resolution (Rs = 0.9) relative to HC (Rs = 0.7) and HC (Rs = 0.5) (Figure 2).

A MeOH/H₂O gradient (10% → 7% water) further increased DHA–DPA separation (best Rs = 1.4) with the HP column (Figure 2).

Scale-up required a step change in loading: analytical-format capacity (30 g class columns; 150 µL oil) was incompatible with producing tens of grams of DPA within reasonable instrument time.

We moved to a two-column flash configuration, a high-performance C18 (15 µm, 330 g) paired with a higher-capacity C18 (45 µm, 330 g), to increase injected mass while maintaining usable selectivity (Figure 3).

Relative to the HP column alone, the coupled system increased the purity of the collected DPA-enriched fraction by approximately 3.5-fold at the tested loading, while run time increased by approximately twofold (Figure 2A; see also Supplementary Table S1 for the numerical values). Among ten MeOH/H₂O programs evaluated (90/10 to 99/1), we selected an isocratic 92/8 mode as the best compromise between throughput, robustness, and per-run recovery (Figure 2).

Under these preparative conditions, per-cycle recovery of DPA in the reinjected central fraction was 91–95% (Supplementary Table S1). Because coupling increased backpressure, we fixed flow at 120 mL/min, corresponding to the maximum pressure tolerated by the columns while retaining efficient cycle times (Figure 3).

3.3. Purification of DPA and n-3 LCPUFAs

We implemented a closed-loop, reinjection-based workflow in which each overloaded flash run is fractionated, analytically triaged, pooled by composition, and reinjected to progressively sharpen the DPA-enriched band (Figure 4).

After equilibration, a 20 mL aliquot of the prepared 4020EE feed solution was separated on the two-column system, and fractions were collected every 30 s (≈60 mL) (Figure 4A).

Each fraction was evaporated, weighed, and analyzed by GC to map DPA purity versus retention time, enabling objective definition of collection windows (Figure 4B).

We then partitioned the chromatogram into four operational pools: a DHA-rich region (>85% DHA), two mixed transition regions (<50% DPA), and a DPA-enriched region (>50% DPA) designated for reinjection (Figure 4B).

Successive injections accumulated material into these predefined pools until sufficient mass was reached, at which point pooled fractions were reinjected to iterate the enrichment cycle (Figure 4C,D).

After three cycles, pooling and reinjection were continued until n-3 DPA reached >98 area% of total identified FAME by GC–FID, enabling production of several tens of grams of highly purified DPA ethyl ester (Figure 5).

To accelerate throughput, the first two cycles were run at room temperature with MeOH/H₂O 99/1, prioritizing rapid removal of EPA (45% of 4020EE) and partial DHA depletion rather than full DHA–DPA resolution (Figure 5).

Consistent with that design, DPA purity increased only modestly (7% → 18%) after these two fast cycles, but the step sharply reduced the volume, requiring higher-resolution processing from cycle three onward (Figure 5).

Subsequent cycles used the optimized 92/8 isocratic mixture to restore selectivity while maintaining preparative throughput (Figure 2).

A remaining bottleneck was the n-3 DPA/n-6 DPA pair, which could not be resolved on the flash system under any tested conditions.

To recover n-3 DPA from the terminal mixed fraction (approximately 60% n-3 DPA and 40% n-6 DPA), we subjected this pool to an additional polishing step and collected the resolved n-3 DPA fraction, which was then combined with the main DPA pool. Under this integrated scheme, the cumulative recovery of n-3 DPA from the starting DPA content reached 76% (Figure 5C). For clarity, the 91–95% value refers to recovery in an individual purification cycle, whereas 76% refers to the integrated overall yield after all enrichment and polishing steps.

3.4. Laboratory Workflow Management and Limitations

We selected fatty acid ethyl esters as the purification target because they can be directly incorporated into semi-synthetic diets and reflect the dominant commercial form of enriched n-3 LCPUFA oils, improving accessibility and cost realism for in vivo programs. Newer market oils enriched in EPA + DPA with minimal DHA could compress cycle count and solvent burden under this workflow.

Although gradients improved DHA–DPA resolution at small scale, we retained isocratic operation at preparative scale to eliminate re-equilibration overhead and to support successive injections as a practical productivity lever.

A practical limitation was the progressive loss of flash-column performance during successive injections, most likely due to retained-material accumulation and the lack of true reverse-flush cleaning on this platform (Figure 4C). Although we did not formally quantify the exact number of injections required before performance declined, the first column functioned as a sacrificial capture bed, helping preserve the resolving performance of the second (15 μm) column. In this program, processing 1.8 L of oil required 8500 L methanol, of which approximately 70% was recovered by distillation. Accordingly, the method should be viewed as scalable for gram-to-tens-of-grams production in a research setting, but not as a direct substitute for optimized industrial high-throughput DPA manufacturing. In addition, ethanol volumes for water removal are needed, alongside high-flow solvent handling (120 mL/min) and solvent-vapor safety controls [12,13].

4. Discussion

n-3 Docosapentaenoic acid (DPA) interconverts with EPA and DHA but also contributes to a dedicated pool of bioactive lipid mediators, including DPA-derived protectins and resolvins linked to inflammation resolution in mucosal tissues [1,2,3,4,17]. This combination of biological relevance and limited commercial availability has made access to well-defined DPA a practical bottleneck for mechanistic and translational studies.

Our results show that this bottleneck can be reduced with widely available reversed-phase equipment by treating preparative LC as an iterative enrichment problem rather than a single-pass separation. The stacked C18 configuration provides a pragmatic balance between resolving power and loading capacity, while the isocratic 92:8 MeOH/H₂O system simplifies re-equilibration and supports routine solvent recovery. In this context, the key design principle is the purification loop: only the central, DPA-rich window is reinjected, progressively compressing the compositional space until DHA and EPA are effectively excluded from the circulating pool. The approach is therefore robust to moderate peak overload and day-to-day variability, because enrichment depends on repeated selection of the same chromatographic region. From a physicochemical perspective, the main challenge is that DPA is separated not from unrelated lipids but from closely neighboring long-chain polyunsaturated species whose hydrophobicity and unsaturation differ only incrementally. Under overloaded conditions, such molecules are expected to exhibit broader bands, partial overlap, and stronger sensitivity to solvent-strength mismatch than would be predicted from dilute analytical separations alone. The success of the present workflow therefore depends less on achieving baseline resolution in a single pass and more on repeatedly isolating the center of the DPA-rich band, where the local compositional variance is progressively reduced across cycles.

Compared with earlier laboratory-scale protocols that produced DPA as an enriched fraction or required specialized stationary phases, the present workflow offers a different trade-off. Argentated systems offer higher isomeric selectivity than standard C18 media, and industrial multicolumn HPLC can deliver far greater throughput than the present workflow. Counter-current chromatography also avoids solid stationary-phase degradation and may be advantageous for some highly unsaturated targets. The advantage of the present method is that it lowers the equipment barrier for academic laboratories by using commercially available flash hardware and a reproducible reinjection logic, albeit at the cost of higher solvent consumption, more hands-on fraction handling, and lower industrial relevance. Also, instrument access and biphasic solvent development remain barriers in many laboratories. Our data show that for most academic settings, repeated isocratic C18 cycles can reach a purity threshold (>98%) that is sufficient for analytical validation and downstream biology, with the remaining limitation being partial co-elution of the n-6 DPA isomer at the tail of the DPA band. Another relevant consideration is that highly unsaturated fatty acid esters are not only chromatographic analytes but also chemically labile amphiphilic molecules. Concentration-dependent association, precipitation of less-soluble components during cold handling, and early-stage oxidation can all distort preparative behavior by altering the effective sample composition before or during separation. In our workflow, low-temperature pretreatment likely removed poorly soluble or oxidized material that would otherwise broaden peaks and contaminate later fractions, while antioxidant protection and short residence times minimized secondary degradation during repeated cycling.

Finally, oxidation control is not a secondary detail but a defining requirement for any multi-cycle PUFA isolation. The low peroxide and p-anisidine values, together with the absence of detectable TBARS in the processed material, indicate that antioxidant protection, cold handling, and short residence times can preserve chemical integrity even over repeated injections. Incorporating solvent recovery further improves practicality and aligns the workflow with emerging expectations for greener preparative chromatography [12,13].

5. Conclusions

In summary, the fundamental contribution of this work is to show that for low-abundance polyunsaturated targets such as n-3 DPA, preparative purification can be redesigned as an iterative compositional-selection problem rather than a single-step separation problem. This work provides a compact, reproducible route to high-purity n-3 DPA ethyl ester at the gram-to-tens-of-grams scale using an iterative, windowed chromatographic enrichment approach. The strategy is operationally straightforward and compatible with standard laboratory chromatography and should be adaptable to other low-abundance LCPUFAs and related lipid targets where neighboring species partially overlap, improving access to purified material for downstream analytical, formulation, and biological studies.