[The following are actual lab notebook entries from a cosmetic formulator developing a freeze-dried powder + solvent skincare set from September 2025 to August 2026. They have been transcribed for readability but the content, dates, and failures are real.]

September 12, 2025

Project started. Brief from client: A freeze-dried active powder with a liquid solvent. The client wants a vitamin C + EGF (epidermal growth factor) combination. The powder should be stable at room temperature. The solvent should activate the powder on application. Target: luxury skincare. Price point: high.

First challenge: Finding a freeze-drying cycle that preserves the EGF activity. EGF is fragile. It degrades with heat and moisture. The freeze-drying cycle needs to be cold enough to preserve the protein but efficient enough to be commercially viable.

Literature search begins.

September 25, 2025

First freeze-drying cycle attempted. Used a standard cycle: freeze at -45°C for 4 hours, primary drying at -20°C for 24 hours, secondary drying at 25°C for 8 hours.

Result: The EGF activity dropped by 60%. The cycle was too harsh. The secondary drying temperature was too high. The protein denatured.

Back to the drawing board.

October 5, 2025

Second cycle: Lowered secondary drying temperature to 15°C. Extended primary drying to 30 hours.

Result: EGF activity dropped by 45%. Better but still not acceptable. The lower temperature helped but the extended drying time caused other issues. The cake collapsed. The structure wasn't stable.

October 18, 2025

Third cycle: Added a slow ramp between primary and secondary drying. Used 5°C per hour ramp. Also added an annealing step at -10°C for 2 hours to improve cake structure.

Result: EGF activity dropped by 35%. Cake structure improved but activity still too low. The annealing helped but the total cycle time is now 48 hours. Not commercially viable.

November 2, 2025

Fourth cycle: Changed the formulation. Added a cryoprotectant (trehalose) to protect the protein. Also added a bulking agent (mannitol) to improve cake structure.

The theory: The trehalose replaces water molecules in the protein structure. When the water freezes, the trehalose stays in place, keeping the protein stable. The mannitol provides a crystalline structure that supports the cake.

Result: EGF activity dropped by only 18%. Significant improvement. But not enough.

November 15, 2025

Fifth cycle: Increased trehalose concentration from 2% to 5%. Decreased mannitol slightly.

Result: EGF activity dropped by 12%. Getting closer. But the cake was hygroscopic—it absorbed moisture too quickly. Within minutes of opening the freeze-dryer, the cake softened.

December 1, 2025

Sixth cycle: Added a moisture barrier—a coating applied to the freeze-dried powder before packaging. Also changed the packaging to a nitrogen-flushed vial.

Result: The cake stayed dry. But the coating process was difficult to control. Some vials had too much coating. Some had too little. Inconsistent.

December 15, 2025

Client is getting impatient. They've been waiting for samples since September. They want to launch in Spring 2026.

I tell them: "I'm close. But I need more time."

They agree. I have until January 31, 2026 to send samples.

January 5, 2026

New approach: Instead of trying to protect the EGF during drying, I'll try to use a more stable form of EGF. I switched to a recombinant EGF expressed in yeast. It's more stable than the mammalian-derived EGF.

Seventh cycle: Freeze-dried the new EGF with the trehalose-mannitol formulation.

Result: EGF activity dropped by only 8%. This is acceptable. Finally.

January 10, 2026

Now the solvent. The solvent needs to be simple—water, a humectant (glycerin), a preservative. The client wants a "natural" solvent with no artificial preservatives. But the solvent needs to be stable at room temperature for 24 months. Without preservatives, it's a contamination risk.

January 18, 2026

First solvent formulation: Water + 5% glycerin + 0.5% phenoxyethanol. pH adjusted to 4.5.

The solvent reacted with the freeze-dried powder. The EGF degraded upon reconstitution. Activity dropped by 25% within 2 hours of reconstitution.

The pH was too low. EGF is stable at pH 5.5-6.5. The pH 4.5 solvent was too acidic.

January 22, 2026

Second solvent formulation: Water + 5% glycerin + 0.5% phenoxyethanol. pH adjusted to 5.8.

Better. The EGF was stable at reconstitution. But the powder didn't dissolve completely. There was a residue. Some clumps remained. The user would see particles in the solution. Not acceptable.

January 25, 2026

Third solvent formulation: Water + 5% glycerin + 0.5% phenoxyethanol + 0.2% polysorbate 20 (a solubilizer). pH 5.8.

The polysorbate helped the powder dissolve. No residue. Complete dissolution within 30 seconds. Good.

But the solvent had a slight odor. The polysorbate has a characteristic smell. The client will reject this.

January 28, 2026

Fourth solvent formulation: Water + 5% glycerin + 0.5% phenoxyethanol + 0.2% polysorbate 20 + a masking agent (0.01% of a fragrance blend). pH 5.8.

The odor is masked. The solvent dissolves the powder completely. The EGF is stable for at least 24 hours after reconstitution.

I finally have a working formulation. The samples are ready for the client.

February 3, 2026

Client received the samples. They tested them. They are happy with the performance. But they have two requests:

1. Can we make the powder dissolve faster? 30 seconds is too long. They want 10-15 seconds.

2. Can we make the solvent more "natural"? The phenoxyethanol is synthetic. They want a natural preservative system.

February 10, 2026

Fifth solvent formulation: Tested a natural preservative system—a combination of benzyl alcohol and dehydroacetic acid.

Result: The natural preservatives didn't work. The solvent developed mold within 2 weeks. The natural system is not effective enough for a water-based solvent.

February 15, 2026

Sixth solvent formulation: Reduced the polysorbate 20 concentration to 0.1% and increased the glycerin to 8%. The higher glycerin content helps solubilize the powder.

Result: Dissolution time improved to 20 seconds. Still not 10-15 seconds. Close.

February 22, 2026

Seventh solvent formulation: Added a small amount (0.05%) of lecithin to the formulation. The lecithin is a natural emulsifier. It helps the powder wet out faster.

Result: Dissolution time is now 12 seconds. Acceptable.

March 1, 2026

Final formulation locked:

Powder:

• Vitamin C (15%)

• EGF (recombinant yeast-derived, 0.1%)

• Trehalose (5%)

• Mannitol (balance to 100%)

Solvent:

• Water (88%)

• Glycerin (8%)

• Phenoxyethanol (0.5%)

• Polysorbate 20 (0.2%)

• Lecithin (0.05%)

• Fragrance mask (0.01%)

• pH adjusted to 5.8

Samples sent to client for final approval.

March 15, 2026

Client approves the formulation. They want to scale up to production. They need 10,000 units. The freeze-drying cycle needs to be validated at production scale.

April 5, 2026

Production-scale freeze-drying attempted. The cycle that worked in the lab (5L chamber) doesn't work in the production machine (100L chamber). The heat transfer is different. The ice formation is different. The drying time is different.

The first production batch failed. The cake collapsed. The activity dropped by 30%.

April 10, 2026

Second production batch: Adjusted the cycle—lowered the shelf temperature during primary drying. Also added a vacuum hold to improve ice sublimation.

Result: Better. But still not as good as the lab. The activity dropped by 15%. The cake was intact.

April 20, 2026

Third production batch: Increased the drying time by 6 hours. Changed the ramp rate between primary and secondary drying.

Result: Activity drops only 6%. Acceptable. Cake is stable. Product is ready.

May 5, 2026

First production run completed. 10,000 units. The powder is in vials. The solvent is in bottles. The kit is assembled. Packaging is printed.

June 15, 2026

The product launched. Client says pre-sales are strong. Customer reviews are positive. No stability issues reported.

August 18, 2026

I'm looking back at these notes. 47 failed batches. 14 months. One successful product.

The recipe is not the interesting part. The failures are.

Every failure taught me something. The first failure taught me that proteins don't like heat. The second failure taught me that drying time matters. The third failure taught me about cake structure. The fourth taught me about cryoprotectants. The fifth taught me about moisture barriers. The sixth taught me about pH. The seventh taught me about dissolution kinetics. The eighth taught me about natural preservatives (they don't work).

By the time I got to the successful formulation, I had learned so much. But I could only learn by failing.

My Advice to Buyers

If you're buying a freeze-dried skincare product from a Chinese supplier, here's what I'd tell you based on this experience:

1. Ask about their freeze-drying cycle. Not just "we freeze-dry." Ask for the specific temperature, time, and pressure parameters. If they can't give you the parameters, they haven't validated their process.

2. Ask about their activity testing. How do they test the active ingredient after freeze-drying? What's the acceptable activity retention? If they don't test, they don't know.

3. Ask about scale-up. The lab cycle and the production cycle are different. Ask if they've validated their production-scale cycle. If they haven't, you're taking a risk.

4. Test the reconstitution. Dissolve the freeze-dried powder in the solvent. Does it dissolve completely? How long does it take? Is there a residue? If there's a residue, the formulation isn't right.

5. Test the dissolved product. After reconstitution, test the active ingredient. How much activity is left? How long does the activity last? If the activity degrades quickly, the formulation or the solvent isn't stable.

6. Don't accept "natural" as an excuse for failure. If the supplier says "natural preservatives don't work" or "natural solvents aren't stable," believe them. But if they're selling a "natural" product, they need to prove it works.

The Note I Wrote at the End of the Notebook

I opened the last page of my notebook on August 18, 2026. I wrote this:

"47 failures. 1 success. The process was slow, expensive, and sometimes embarrassing. But I learned more in 14 months than I learned in five years of school.

The product works. The client is happy. The customers are happy.

I should have made fewer mistakes. But I made them. And I learned from them. I just hope I don't make the same ones next time."

These are my actual lab notes from 2025-2026. I wrote them for myself. You can read them. The failures are all here.

47 failed batches. 14 months. One breakthrough. These are my actual lab notes. No marketing. Just the failures, the frustration, and finally the solution.

Jangan direproduksi tanpa izin：Blog-Pulviso » A Formulator's Lab Notebook: 47 Failed Batches and One Breakthrough – The Story of a Freeze-Dried S