Draft, not validated, not approved. Content is user-generated and unverified. Every number marked “proposed” or “inference” is a starting assumption, not a sourced standard. Verify against the cited references and your own regulator before use.
CONTROLLED DOCUMENT · DRAFT, NOT VALIDATED, NOT APPROVED
Pesticide remediation of mother liquor, lot OBC-CRD-072026-002-PRDIST
Document no.
SOP-OBC-CRD-072026-002-PRDIST-R0
Revision
R0, draft
Source COA
Encore Labs, 2607ENC0296_1019
COA method
SOP CA_M-PESTMYCOLCMS
COA tested
08/03/2026
Producer
MoloWorx, AL
Jurisdiction
Alabama (ADAI, ADEM)
Target
non-detect at the lab LOD
1. Purpose and scope
Remove the residues listed in section 2 from mother liquor lot OBC-CRD-072026-002-PRDIST to non-detect, retaining as much cannabinoid as the chemistry allows, and prove the removal by the checks in section 4. The material is viable product. Waste handling applies only to removed fractions and used adsorbent.
Absolute zero cannot be measured. "Removed" in this SOP means non-detect at the reporting lab's LOD, on a method shown to detect the analyte in the final product matrix.
2. Basis: what the COA found
Analyte
Result µg/g
Limit
LOD
LOQ
Target
Cut needed
Myclobutanil
0.134
0.100
0.011
0.033
0.011
12×
Boscalid
0.045
0.100
0.01
0.031
0.01
4.5×
Bifenthrin
1.03
3.00
0.049
0.150
0.049
21×
Permethrin
0.155
0.500
0.013
0.039
0.013
12×
Bifenazate<LOQ
0.03
0.100
0.01
0.03
0.01
3.0×
Propiconazole
0.029
0.100
0.009
0.028
0.009
3.2×
Acetamiprid
0.027
0.100
0.006
0.018
0.006
4.5×
Azoxystrobin<LOQ
0.017
0.100
0.006
0.017
0.006
2.8×
Metalaxyl<LOQ
0.021
2.00
0.007
0.021
0.007
3.0×
Limits are those printed on the COA by a California lab for an Alabama producer. They are not Alabama limits. A result of <LOQ is treated as its LOQ for planning, and does not count as non-detect. This COA is for distillate; the mother liquor itself must be assayed to set the true baseline.
Lot-specific risks
Over the limit: Myclobutanil (1.34× limit).
No removal data found for Boscalid, Propiconazole, Acetamiprid, Azoxystrobin and Metalaxyl. Behavior on the column is an assumption until Phase 0.
Bifenthrin: May tail into later fractions because of strong sorption.
Permethrin: Melting point 34.5 C, so it can soften near a 35 to 40 C dissolve or dry step.
Permethrin: Report both isomers.
Bifenazate: Fast hydrolysis: keep water out and the pH neutral.
9 analytes at once usually points to a source or input problem. Trace the biomass lot before spending on cleanup.
3. Definitions
ND: result below the LOD on the COA. <LOQ: detected but not quantifiable, and not ND.
Reduction factor: baseline µg/g ÷ final µg/g. If the final is ND, it is reported as at least baseline ÷ LOD.
Pool: fractions combined as product. Shoulder: cannabinoid-rich fractions just outside the pool, held for a second pass.
Closure: analyte mass recovered across all streams ÷ analyte mass loaded.
4. Acceptance criteria and verification
Removal is verified four ways. Each has a measurable criterion. Ranges marked proposed are my proposals from common analytical practice, not sourced standards, so set them in your validation policy.
4.1 Release criteria for the final product
#
Criterion
Measured by
Accept when
A1
Every target analyte is not detected
Same LC-MS/MS panel and method as the COA
ND (below LOD) for each analyte in 4.2
A2
The method can see the analyte in this matrix
Matrix spike of each analyte into the final product at its LOQ
Recovery 70 to 120% proposed
A3
No contamination from the process
Process blank (solvent through the column) and a carryover blank between lots
All target analytes ND
A4
Removal, not loss
Mass balance, section 4.3
Closure 80 to 120% per analyte on validation batches proposed
A5
Degradation did not fake removal
Assay for known degradates where they apply (bifenazate degradates: none identified here, ask the lab)
Degradates ND, or, if not assayable, closure in A4 holds
A6
Independent confirmation
Split final sample to a second accredited lab (ADAI-accepted, confirm with ADAI)
Same ND result
A7
Every other analyte on the panel passes
Full COA panel
All Pass
A8
Cannabinoid retention
Potency by HPLC on every fraction; mass in = pool + shoulder + waste
Recovery at or above your set target: and mass balance 90 to 110% proposed
4.2 Release results, per analyte
Analyte
Baseline µg/g
LOD
Final result
Matrix spike recovery %
Second lab
Pass
Myclobutanil
0.134
0.011
Boscalid
0.045
0.01
Bifenthrin
1.03
0.049
Permethrin
0.155
0.013
Bifenazate
0.03
0.01
Propiconazole
0.029
0.009
Acetamiprid
0.027
0.006
Azoxystrobin
0.017
0.006
Metalaxyl
0.021
0.007
4.3 Mass balance, proof that the molecules went somewhere else
On Phase 0 and each validation batch, assay every fraction, the column strip (strong solvent wash of the used adsorbent) and the rinses. Enter analyte mass in µg (ng/mL × mL ÷ 1000). A closure far below 80% means the analyte was lost or degraded, and the removal claim fails.
Baseline column defaults to the COA value. Overwrite it with the mother liquor assay.
Analyte
Baseline µg/g
In µg
Pool µg
Shoulder µg
Other fractions µg
Column strip µg
Out µg
Closure
Final µg/g or ND
Reduction
Myclobutanil
–
–
–
–
Boscalid
–
–
–
–
Bifenthrin
–
–
–
–
Permethrin
–
–
–
–
Bifenazate
–
–
–
–
Propiconazole
–
–
–
–
Acetamiprid
–
–
–
–
Azoxystrobin
–
–
–
–
Metalaxyl
–
–
–
–
4.4 Process validation and what it proves
Run three consecutive validation batches on this lot type. Each must meet A1 to A8. proposed
Three passes in a row prove repeatability on those three runs. They do not prove a pass rate: with zero failures in 3 runs, one can claim only 95% confidence that at least 37% of batches pass. Claiming 90% reliability needs 29 passing runs, and 95% needs 59.
So release is by lot, not by process. Every production lot is tested to A1 to A3 and A6 before it leaves quarantine, until ADAI says otherwise.
Any change to the adsorbent lot, solvent lot, loading, flow, feed matrix or column size reopens validation for that change.
4.5 Out of specification
Any failed criterion puts the lot on hold and is logged as a deviation.
Do not reblend the lot or rerun it without a written plan approved by the reviewer. A second pass is allowed only under a pre-approved protocol with its own mass balance.
If the analyte appears in the pool, change the adsorbent lot and repeat Phase 0 before the next batch.
5. Retention of cannabinoids
Estimated first-pass recovery under the prior model is 22% (P10 to P90 0 to 82%). This is assumption-driven and is replaced by the Phase 0 optimizer result once fraction data exists.
The pool is the widest cannabinoid block that meets every target, chosen from fraction data.
Shoulder fractions are held and re-run. Activated carbon and distillation are not used as the fix.
Record every fraction's cannabinoid mass. Pooled cut points:
6. Procedure
Process values to record after Phase 0: loading concentration adsorbent per g solids flow fraction volume
Phase 0: Bench validation for this lot
Nothing here is validated. This phase produces the numbers that replace every assumption on this page.
0.1Baseline. This COA is a distillate result. Send a retained sample of the mother liquor itself for the same panel and method (SOP CA_M-PESTMYCOLCMS), plus cannabinoid potency. I expect mother liquor to carry more residue than distillate, but that is inference.
0.2Blanks. Run solvent and glassware blanks through the whole procedure and resolve any peak first. Avoid plastics: phthalates leach.EPA 8080
0.3Spike clean isolate in hexane with the target analytes at levels the lab can quantify in a fraction (at least 10 times the LOQ): Myclobutanil 0.330 µg/g; Boscalid 0.310 µg/g; Bifenthrin 1.50 µg/g; Permethrin 0.390 µg/g; Bifenazate 0.300 µg/g; Propiconazole 0.280 µg/g; Acetamiprid 0.180 µg/g; Azoxystrobin 0.170 µg/g; Metalaxyl 0.210 µg/g. Keep everything near neutral pH and ambient temperature.
0.4Normal-phase map. Silica at about 10% sample-to-media ratio, and separately Florisil, both loaded in hexane. Step up eluent polarity in small increments and collect small numbered fractions.EPA 3620C, Biotage
0.5Reverse-phase map. Run C18 with a neutral ethanol and water gradient at about 3% loading for the moderately polar analytes (Myclobutanil, Propiconazole, Acetamiprid and Azoxystrobin). Do not add acid.Biotage
0.6Analyze every fraction by the same LC-MS/MS panel that produced this COA for Myclobutanil, Boscalid, Bifenthrin, Permethrin, Bifenazate, Propiconazole, Acetamiprid, Azoxystrobin and Metalaxyl, and by HPLC for cannabinoids. Same method means the fraction numbers compare directly to the COA.
0.7Enter the results in the Phase 0 optimizer below. It picks the widest cannabinoid pool that meets every target. Target for this lot: Myclobutanil < 0.011; Boscalid < 0.01; Bifenthrin < 0.049; Permethrin < 0.013; Bifenazate < 0.01; Propiconazole < 0.009; Acetamiprid < 0.006; Azoxystrobin < 0.006; Metalaxyl < 0.007 µg/g.
0.8Capacity test. Raise the mass of mother liquor per gram of adsorbent until the first analyte appears in the pooled cannabinoid fractions. Set production loading below that point.
0.9Go / no-go. Proceed only if the optimizer finds a pool that meets every target and the recovery is acceptable to you. If not, use the fallback under Retention.
0.10Permethrin. Read cis and trans separately. If it overlaps the cannabinoid peak on both silica and Florisil, do not spend a production batch on it. Its C18 record is poor.J Cannabis Res
Phase 1: Pretreatment
Fill the orange values from Phase 0.
1.1Record mother liquor lot, source biomass lot, mass, volume and cannabinoid concentration. Keep a sealed sample.
1.2Winterize if waxes or fats are present, to keep the column from clogging.Root Sciences
1.3Exchange to {conc} in pesticide-quality hexane. Hexane as loading solvent is required by the EPA Florisil method. Filter through glass fiber or PTFE.Method 3620A
Phase 2: Column run and pool
The pool is chosen from data, not by a fixed rule.
2.1Pack {load} g activated adsorbent per gram of solids under anhydrous sodium sulfate. Pre-elute with hexane. Run at {flow} mL/min.
2.2Load without letting the bed run dry. Elute the validated steps in {fvol} mL fractions.
2.3Assay every fraction for cannabinoids and send the boundary fractions for the target analytes before pooling. Hold everything in labeled glass.
2.4Pool the fractions the optimizer selected. Put the shoulder fractions (cannabinoid-rich but just outside the pool) in a separate labeled container for a second pass. That is where the yield is recovered.
2.5Move analyte-bearing fractions and used media to labeled hazardous-waste containers pending the ADEM waste determination. Never into general solvent waste.
Phase 3: Recover and crystallize
Treat every new mother liquor as untested.
3.1Filter the pool through 0.45 µm PTFE, then strip solvent on a rotary evaporator at a low bath temperature (heat degrades the labile analytes). Label recovered solvent as untested.
3.2Recrystallize from petroleum ether at about 1.5 to 3 mL per gram of cannabinoid, dissolve near 40 C (permethrin melts at 34.5 C, so it can soften at this step), cool to 4 to 23 C for hours, vacuum filter, wash cold, dry under vacuum.US 9,765,000, US 10,239,808
3.3Do not fold the new mother liquor into another batch until it has a result. Recycling concentrates impurities, and I expect it would concentrate residues too (inference).B/R Instrument
3.4Quarantine the lot until release testing is complete.
Release: QC and acceptance
Only the lab result counts. Never release on an in-process number.
R.1Send the final product for the full panel at Encore Labs or an Alabama-accepted lab (confirm with ADAI). Same method as the baseline.
R.2Accept when every target analyte is non-detect at its LOD, or below the target in the table above, and every other analyte on the panel passes. Fail analytes for this lot: Myclobutanil.
R.3Mass balance. Record cannabinoid mass in, pooled, shoulder and waste for every run. Investigate any unexplained loss.
R.4Spike recovery on one aliquot per 10 batches or each new adsorbent lot.
R.5A failed release puts the lot on hold. Do not reblend without regulatory guidance.
7. Records
Batch record: mother liquor lot, source biomass lot, mass, volume, adsorbent type and lot, solvent lots, loading, flow, fraction volumes and results.
Operator, reviewer, date and time for each phase, and every deviation.
Retain a sealed sample of the mother liquor, the pool and the final product.
8. Waste and safety
Segregate three streams: analyte-bearing fractions, used adsorbent and filters, and recovered solvent. Label each with contents, lot and date.
Run the ADEM waste determination (Ala. Admin. Code r. 335-14-3-.01(2)) on the fractions and spent media before shipment. Waste codes were not checked for Myclobutanil, Boscalid, Bifenthrin, Permethrin, Bifenazate, Propiconazole, Acetamiprid, Azoxystrobin and Metalaxyl. The solvent stream is ignitable (D001) if hexane or petroleum ether.
Solvent handling follows your extraction-room controls. Use nitrile over laminate gloves, splash goggles, lab coat and a fume hood. General practice, not from a source.
9. Open regulatory items
Alabama limit for each analyte above, and whether sale of remediated material from a contaminated lot is allowed.
Which labs ADAI accepts for release testing, and whether your ADAI license covers this step.
Which agency governs finished CBD products. One secondary source names the ABC Board under HB 445, so verify against the statute.
Your ADEM generator category and any notification duty.