Barrier Repair Formula Spec: Ceramide, PGA, Panthenol Targets for Indie K-Beauty (2026)

By the ALTA MEET editorial team — K-beauty ODM consulting

Barrier repair is the highest-CTR long-form content on this site. Founders keep returning to it because the “compromised barrier” complaint is the single most common brief we hear from an indie K-beauty founder building a first hero product. What is missing from most of the barrier-repair conversation, though, is the piece a Korean ODM actually needs to price and formulate a product: a spec sheet. Not a hero copy line, not a marketing angle, but a list of three or four molecules, target concentrations, molecular-weight preferences, source form, and stability constraints.

This post walks through a founder-grade spec sheet built around three of the most-cited barrier actives in the peer-reviewed literature and in Korean ODM formulation practice: ceramides, polyglutamic acid, and panthenol. Every concentration and structural claim below is grounded in a public source or the EU CosIng ingredient database, and each section closes with the specific line an indie founder should put on the RFQ so that quotes from Korean labs come back on the same scale.

Why these three actives, and why together

The stratum corneum barrier is a lipid–protein composite. The lipid matrix is composed of roughly equimolar amounts of ceramides, cholesterol, and free fatty acids arranged in short and long periodicity lamellar phases with repeat distances of about 6 and 13 nanometers respectively (PMC3117011). Ceramides represent roughly half of the SC lipid mass and are classified into twelve subclasses in human skin (PMC7461267). When the ratio of these lipids is disturbed, or when free fatty acid or ceramide content drops, the lamellar organization breaks down and transepidermal water loss rises. That is the physical event underneath “my skin feels tight, red, and reactive.”

The three actives on this spec sheet each act on a different lever in that system:

Ceramide is the direct structural replacement. Topical application of physiological ceramides interacts with the SC lipid matrix and, in ex vivo compromised skin, integrates into the lamellar phases (PMC5801391). A qualitative review of ceramide-containing formulations shows that hydration and TEWL reduction persist for roughly twenty-four to forty-eight hours after application, versus four to eight hours for formulations without physiological lipids (Kono 2021, J Dermatology, DOI 10.1111/1346-8138.16175).

Polyglutamic acid, specifically the gamma isoform produced by Bacillus fermentation, is a high-molecular-weight polyanionic humectant that pulls water into the upper stratum corneum and, at least in keratinocyte and reconstructed skin models, upregulates barrier-marker genes and aquaporin-3 (Kwon 2025, Int J Mol Sci, DOI 10.3390/ijms26030983). Reported molecular weights for cosmetic-grade γ-PGA range from about 100 to more than 1000 kilodaltons (PubMed 29229248), and the MW you specify changes both the sensory profile and how deeply it deposits.

Panthenol, the alcohol form of vitamin B5, is a small-molecule humectant that gets metabolized to pantothenic acid on skin. Semisolid formulations containing seven to eleven weight percent panthenol have shown increased skin hydration and decreased transepidermal water loss in vivo (Camargo 2018, PubMed 29577586), and TEWL inhibition is reported to appear at concentrations of about one percent and above. In a double-blind randomized study of a panthenol-enriched mask used after facial laser treatment, TEWL was significantly reduced at day 3, day 7, and day 14 versus control (Gao 2025, J Cosmet Dermatol, DOI 10.1111/jocd.70223).

Together, this trio covers three distinct mechanisms: structural replacement of a depleted lipid, polymer-scale water binding at the surface, and a small-molecule TEWL suppressor with a soothing profile. That is why a barrier repair product often reads better than a barrier repair product built on any one of them alone. A recent clinical study on sensitive skin used a combination of Centella asiatica extract, Ceramide NP, and panthenol and reported improvements in hydration and barrier metrics over four weeks (PMC12274965).

Prerequisites before writing the RFQ line

Before a Korean ODM can price this formula, three decisions have to be closed on the founder side.

Skin condition target. Everyday dry sensitive skin, atopic-adjacent skin, and post-procedure or laser-recovery skin all pull the spec in slightly different directions. Post-procedure skin usually pushes panthenol higher and adds a soothing botanical; atopic-adjacent skin tends to push ceramide subclass diversity up and add cholesterol.

Product format. Cream, ampoule, gel, and sheet mask each impose different upper limits on ceramide loading because ceramides are lipophilic solids with limited water solubility. In a light aqueous ampoule the practical ceramide ceiling is much lower than in a rich oil-in-water cream. Delivery-system work such as nanoemulsions was developed precisely because raw ceramides have low permeability into the SC (PMC12108929).

Regulatory geography. Korea, the US, and the EU allow all three actives in leave-on cosmetic products, but the compliance overhead differs. Ceramide NP, Ceramide AP, Ceramide NS, Ceramide 1, and Ceramide 3 are all in the EU CosIng database (see for example CERAMIDE AP and CERAMIDE NS). Panthenol is a widely used cosmetic ingredient. Under Korea’s functional cosmetics framework, an indie founder should note that ingredients on the MFDS notified list can be used at their designated concentration without pre-market approval of the finished product, whereas non-notified ingredients trigger a fuller efficacy dossier (ChemLinked, South Korea Functional Cosmetic Regulation). A pure barrier-repair claim is typically framed as a general cosmetic claim rather than a functional cosmetic claim, but a Korean ODM will want the founder to be explicit about which market the product is destined for so packaging INCI and quantity language can be prepared correctly.

Section A: Ceramide selection

The two decisions here are subclass and target concentration.

Subclass. Human SC ceramides are named with a two-letter code that identifies the sphingoid base (S sphingosine, dS dihydrosphingosine, P phytosphingosine, H 6-hydroxysphingosine) and the fatty acid class (N non-hydroxy, A alpha-hydroxy, O omega-hydroxy). Ceramide NP (N-stearoyl phytosphingosine) and Ceramide AP (alpha-hydroxy phytosphingosine) are among the most commonly used cosmetic ceramides because phytosphingosine ceramides are relatively abundant in the human SC and the analogs are commercially available at scale. Ceramide NS and Ceramide EOS (the omega-esterified version) are also frequently specified. The ratio of ceramide NP to NS has itself been proposed as a marker of epidermal differentiation state (PMC7461267), which is one reason a serious barrier formulation will list two or three ceramide subclasses rather than a single one.

Concentration and companion lipids. The classic barrier-restoration work from Man and Elias showed that an equimolar mix of cholesterol, ceramides, and essential and non-essential free fatty acids allowed normal barrier recovery in aged skin, while a 3:1:1:1 ratio (dominant lipid to the other three) accelerated recovery (PubMed 9308554). Practically, that means a founder should not specify a raw ceramide percentage in isolation; the ODM will need to know the intended cholesterol and free fatty acid loading too. A common founder mistake is to ask for “three percent ceramide” without noting cholesterol, which forces the lab to pick a ratio for you.

Delivery form. If the destination format is a thin aqueous ampoule rather than an emollient cream, the RFQ line should request either an oil-in-water nanoemulsion base or a pre-dispersed ceramide complex (many Korean and Japanese raw material suppliers ship “multi-lamellar emulsion” or “pseudo-ceramide complex” premixes) so that solubility limits are moved into the raw material rather than into the finished product.

Spec line for the RFQ: “Physiological ceramide complex, ceramide NP and ceramide AP present at minimum, with cholesterol and free fatty acid loaded at a lipid ratio in the 1:1:1 to 3:1:1:1 range per the Man and Elias barrier repair literature. Please confirm delivery form (raw crystalline versus multi-lamellar emulsion), maximum stable loading in our target base, and whether an additional ceramide subclass (NS or EOS) is included in the supplier premix.”

Section B: Polyglutamic acid selection

The three decisions here are isoform, molecular weight grade, and target concentration.

Isoform. Cosmetic-grade PGA is almost always the gamma (γ) isoform, produced by Bacillus subtilis, Bacillus licheniformis, or a related fermentation. Alpha-PGA does exist but is not what appears on skincare INCI lists. The RFQ should specify “gamma-polyglutamic acid” or the sodium salt “sodium polyglutamate” to remove ambiguity.

Molecular weight. Reported cosmetic and biomedical grades of γ-PGA span roughly 100 to over 1000 kDa (PubMed 29229248), and 400 kDa and 150 kDa are common reference points in the literature (PubMed 25791849). Higher molecular weight grades feel more film-forming on skin and are the ones that have shown barrier-marker upregulation in reconstructed skin models (DOI 10.3390/ijms26030983). Lower molecular weight grades feel lighter and are easier to formulate at higher percentages without an obviously sticky finish. If the format is an ampoule, a mid-to-high MW grade at a low percentage often reads well.

Concentration. The Kwon 2025 keratinocyte and reconstructed skin work reported dose-dependent barrier marker responses at concentrations in the tenths-of-a-percent to low single-digit percent range typical of cosmetic use (DOI 10.3390/ijms26030983). Founders should not chase headline “five percent PGA” claims without checking sensory: the film that gives PGA its barrier feel is the same film that becomes tacky at high loading.

Spec line for the RFQ: “Gamma-polyglutamic acid (or sodium polyglutamate) from Bacillus fermentation. Please confirm molecular weight grade in kilodaltons and share the sensory profile at the top of the supplier’s recommended cosmetic use range. We are targeting a film-forming humectant sensory in a leave-on ampoule format; please note the maximum concentration that maintains a non-tacky finish in your current base.”

Founder Note

I’m Liz, I run altameet from Manhattan, NYC. Barrier repair is the single most-brief we see from indie K-beauty founders, and almost every time the pitch and the RFQ line do not match. The pitch is “calm reactive skin.” The RFQ is “three percent ceramide.” A Korean ODM cannot price the second one on its own. If you want a 15-minute gut-check on whether the spec you are about to send to a Korean lab actually reflects the product you want to launch, write me at liz@altameet.com or book a slot.

Section C: Panthenol selection

Panthenol is the simplest of the three actives to spec, but there are still two decisions to make.

Isomer. D-panthenol (dexpanthenol) is the biologically active isomer. DL-panthenol is a racemic mix and only half of it is active. Cosmetic INCI lists usually just say “panthenol” but a founder should ask the ODM to confirm which isomer is in the raw material and, if it is DL-panthenol, whether the concentration on the label refers to total or to active isomer. This is the single most avoidable mistake on a barrier-repair spec sheet.

Concentration. Published semisolid formulations containing seven to eleven weight percent panthenol increased hydration and decreased TEWL (PubMed 29577586), and TEWL inhibition appears at roughly one percent and above. A leave-on cream aimed at compromised or post-procedure skin will often sit in the 2 to 5 percent range, while a light everyday ampoule commonly sits closer to 1 percent. Post-laser or post-procedure products can go higher, in line with the panthenol-enriched mask study that showed day-3, day-7, and day-14 TEWL reductions (DOI 10.1111/jocd.70223).

Spec line for the RFQ: “D-panthenol (dexpanthenol), with target concentration confirmed against the intended TEWL and hydration benchmarks. If DL-panthenol is used, please quote separately and confirm active-isomer basis on the label. We expect a leave-on cosmetic claim rather than a functional cosmetic claim; please flag if concentration or claim language shifts the regulatory pathway in Korea, the US, or the EU.”

The one-page RFQ template

Sending a Korean ODM a single tidy block of text instead of a scattered email is the difference between a quote in two weeks and a quote in six. Copy this block, fill in the bracketed fields, and paste it below the header of your RFQ email.

Product: [Barrier repair cream / ampoule / mask]
Target user: [Compromised barrier / atopic-adjacent / post-procedure]
Format constraints: [Fragrance-free? Alcohol-free? Fermented preservative?]
Ship-to region: [Korea / US / EU]

Active 1: Physiological ceramide complex
- Ceramide NP + Ceramide AP at minimum
- Cholesterol + free fatty acid at 1:1:1 to 3:1:1:1 lipid ratio
- Delivery form: [Pre-emulsified multi-lamellar / raw crystalline]
- Please state maximum stable loading in your target base

Active 2: Gamma-polyglutamic acid (or sodium polyglutamate)
- Bacillus fermentation
- Please state MW grade in kDa
- Target sensory: film-forming humectant, non-tacky finish
- Please state max loading that maintains sensory in your base

Active 3: D-panthenol (dexpanthenol)
- Confirm isomer; if DL-panthenol, quote separately on active-isomer basis
- Target concentration range: [1 to 5 percent for daily use, higher for post-procedure]

Claims target: leave-on cosmetic, hydration + TEWL reduction
Regulatory pathway: [Korea general cosmetic / US MoCRA / EU]
Stability: 12-month, 45C accelerated + freeze/thaw
Sample turnaround: [4 weeks preferred]
MOQ tolerance: [state realistic first-order units]

Common mistakes on barrier repair spec sheets

Asking for a raw ceramide percentage without ratios. Ceramide-only briefs force the ODM to guess cholesterol and free fatty acid loading. The Man and Elias literature is clear that ratio matters more than absolute concentration for the physical event of barrier restoration.

Specifying a single ceramide subclass. Real SC contains twelve subclasses. A single-subclass raw material is easier to source but less biomimetic. Two or three subclasses (NP + AP, or NP + AP + NS) are a modest sourcing bump for a much more defensible label.

Chasing high PGA percentages for headline claims. The high-MW film that gives PGA its barrier feel becomes tacky at loading levels above the sensory sweet spot. A quiet mid-range percentage with a specified MW is stronger both in feel and in the mechanistic data.

Not distinguishing D from DL panthenol. Halving the active isomer is a silent margin hit and, at low label percentages, can drop actual dexpanthenol below the one-percent TEWL-inhibition threshold reported in the literature.

Locking claim language before the regulatory geography is chosen. A hydration or barrier-support claim is much easier to defend as a leave-on cosmetic claim than as a functional cosmetic claim in Korea. Deciding this after the copy is written is what causes packaging redo cycles.

What to expect over time

Barrier repair studies in the clinical literature typically report measurable improvements on the following cadence, so a founder should benchmark their claim windows accordingly:

Days 1 to 7. TEWL reduction is often the first measurable signal, especially in post-procedure or laser-recovery skin. The Gao 2025 panthenol mask study showed TEWL reduction at day 3 (DOI 10.1111/jocd.70223).

Weeks 2 to 4. Hydration and clinical dryness scores tend to move next. In the ceramide-precursor atopic dermatitis moisturizer study, four weeks of treatment produced significantly greater reductions in TEWL and clinical dryness and greater increases in skin hydration versus control (PubMed 22812593).

Weeks 4 to 12. Sensory endpoints (itch, tightness, reactivity) tend to trail the physical barrier metrics. Founders should push back on marketing pushes for “visible in 3 days” if the endpoint being described is a subjective calm rather than TEWL or corneometer data.

Variations by skin condition

Compromised barrier from over-exfoliation. Ceramide complex is the workhorse. Keep panthenol in the 2 to 5 percent range and PGA at a mid MW grade for a comfortable film. Skip fragrance and dye. This is the highest-volume version of the brief.

Atopic-adjacent daily use. The lipid ratio should lean toward the equimolar-to-3:1:1:1 range that the Man and Elias work showed accelerated barrier recovery (PubMed 9308554). Consider adding a soothing botanical such as centella extract, which is documented in a combined ceramide-NP-plus-panthenol clinical study (PMC12274965).

Post-procedure or post-laser recovery. Panthenol goes higher, up to the 7 to 11 percent range that has published hydration and TEWL data (PubMed 29577586). PGA stays present but at conservative loading to keep the film comfortable on sensitized skin. Ceramide complex remains for structural repair. The Gao 2025 mask study showed TEWL improvements at day 3, 7, and 14 for a panthenol-enriched mask in this population (DOI 10.1111/jocd.70223).

Everyday sensitive skin without a specific insult. The lightest version of the brief. Lower ceramide loading, mid-MW PGA at a comfortable percentage, panthenol closer to 1 percent. This is often paired with a hydrating toner or essence in the routine rather than doing all the work in one product.

Cost and sourcing signals

Detailed unit costs vary by supplier and by market, and any quote a founder receives from a Korean lab will be shaped by three structural drivers rather than by any single ingredient percentage on this list.

The first driver is ceramide subclass and source. Single-subclass, higher-purity ceramide raw materials from Japan or Korea sit at a different price band than blended multi-lamellar emulsion premixes. Asking for two or three subclasses in the same formula raises formula complexity slightly and can push a lab toward a premix option.

The second driver is PGA molecular weight and origin. High-MW Bacillus-fermented γ-PGA is a distinct raw material from mid- or low-MW grades; supply is dominated by a handful of fermentation suppliers, and the choice of MW propagates into both raw material cost and sensory work in the lab.

The third driver is panthenol isomer purity. D-panthenol raw material is typically priced above DL-panthenol on an active-isomer basis, and a founder who specifies dexpanthenol is choosing a slightly higher raw material line for cleaner label language.

None of these three are the largest driver on the total quote. The largest drivers are still formula complexity, fill-line changeover, packaging spec, and MOQ, all of which we have covered in the ODM cost pieces on the site. But the three structural signals above will shift the raw material sub-total inside a barrier repair quote enough that founders should ask about them explicitly.

Key Takeaways

1. A barrier repair brief that lands cleanly at a Korean ODM specifies three actives on three axes rather than one active on a single percentage line.

2. Ceramides are the structural lever. Specify at least two subclasses (NP and AP as a floor, NS or EOS as an extension) and always name the intended cholesterol and free fatty acid ratio alongside the ceramide percentage.

3. Polyglutamic acid is the polymer-scale humectant lever. Specify gamma isoform, molecular weight in kDa, and target sensory rather than chasing headline percentages.

4. Panthenol is the small-molecule TEWL lever. Specify D-panthenol (dexpanthenol) explicitly to protect the active-isomer percentage, and adjust concentration by skin-condition target.

5. Regulatory geography (Korea, US, EU) shapes claim language and packaging INCI. Choose the destination market before locking the copy and the spec.

6. TEWL and hydration are the credible early endpoints; sensory endpoints trail the physical barrier metrics.

FAQ

Do I need all three actives in the same product? No, but a leave-on cream aimed at a compromised or reactive barrier is stronger with all three than with any one alone, because each addresses a different lever (structural, polymer-scale, small-molecule). Ampoules or essences can carry one or two of them and rely on the rest of the routine for the others.

Can I substitute pseudo-ceramide for physiological ceramide? Pseudo-ceramides are patented synthetic analogs designed to mimic the barrier-restoring behavior of physiological ceramides. They are legitimate raw materials but they are not identical to Ceramide NP, AP, or NS. Ask the ODM to disclose which class of ceramide raw material is being used and check that the INCI on your packaging is not misleading.

What molecular weight of PGA should I pick if I do not know the sensory yet? Ask the ODM to bench two options at the top of the supplier’s recommended range, one mid MW and one higher MW, and evaluate on the same base. Sensory is the tie-breaker.

Is 5 percent panthenol overkill for daily use? Not necessarily, especially for the compromised-barrier version of the brief. TEWL inhibition is reported to appear from around 1 percent, and studies of formulations at 7 to 11 percent have shown hydration and TEWL improvements (PubMed 29577586). The trade-off is texture and cost, not safety.

Will Korea classify a barrier repair product as a functional cosmetic? Usually no. Barrier support and hydration claims are generally framed as leave-on cosmetic claims rather than functional cosmetic claims under the MFDS framework, which reserves “functional” status for whitening, anti-wrinkle, and UV protection product categories among others (ChemLinked). Confirm with the ODM on your specific claim language.

Can I move faster by picking a supplier premix that already contains all three? Sometimes yes, especially for a first launch. Premix routes trade optionality for speed, and a founder should still read the INCI, confirm the ceramide subclasses, and ask about the PGA MW and panthenol isomer.

How do I connect this spec sheet to the rest of the launch? This sits on top of the sample-evaluation and RFQ steps we cover elsewhere on the site. Read the Korean ODM sample evaluation protocol next, and reference the how to brief a Korean ODM template for the surrounding email structure.

Working with altameet

altameet is a K-beauty ODM consulting studio based in Manhattan, NYC with Seoul-side partners. If you are trying to translate a barrier repair pitch into a spec sheet a Korean ODM can quote, the fastest way to get a second set of eyes is a 15-minute working session with Liz. Write to liz@altameet.com or book a slot.

Reviewed for accuracy by ALTA MEET’s formulation consulting team.