IQF Broccoli for South Korea: Source Controls, Testing and Bulk Supply Plan

The buyer should treat this as a purchase-order design. It is deliberately narrower than a catalogue claim: no seasoning, no organic designation, no consumer retail label and no ready-to-eat promise. If any of those elements change, we and the importer revisit the specification and Korean compliance route before production. The objective is a repeatable decision on a representative shipment lot, not a perfect-looking hand sample.

Set the Korean buyer floret and pack specification

The first commercial decision is what the buyer will cook or manufacture from the florets. A foodservice kitchen may value uniform crown size for a plated side dish; a soup or ready-meal producer may prefer a different cut and tolerate more small pieces. We would put that choice into a written size band, stem-length limit, maximum broken pieces, colour description, allowed ice or fines, case net weight, inner-liner form and carton dimensions. Those are buyer specifications, not universal Korean statutes. The chosen sampling plan and defect definitions should be included so receiving staff can apply the same standard as the pre-shipment reviewer.

Product identity is a separate legal layer. The SKU remains one ingredient, broccoli, grown and processed in China, blanched and IQF frozen. We would list the factory, production lot, packing date, shelf-life basis, storage instruction and further-cooking statement on a master specification for importer approval. The Korean importer decides the lawful Korean product name and its label route. "Blanched" should not be marketed as "sterilized" or "safe to eat uncooked"; the degree of microbial reduction depends on the actual process, and the intended use here requires cooking. Likewise, an industrial export carton does not resolve a later Korean retail repacking label.

GreenLand IQF broccoli florets for Korean buyer specification.

We would ask for a forecast by month, not only a one-container request. Field availability, size distribution and demand windows affect what portion of a harvest can meet a tight crown specification. A buyer asking for a narrow band, short stems, low defect rate and strong green colour should understand how that affects yield and sorting time. We can discuss an economical manufacturing grade alongside a visual premium grade, but each must have measurable acceptance criteria. "Grade A" has little value if neither party has defined the grade or the receiving sample method.

Before confirming price, the importer checks its imported-food business registration and the actual Chinese factory's registration in MFDS Maru. We supply the manufacturer's legal name and address so the same identity appears on the specification, registration and shipment documents. Maru's current service page lists a three-day processing period and no application fee. Article 5 of the Special Act on Imported Food Safety Control, effective 8 October 2026, requires the registration application before the food-import declaration; its current wording does not impose the seven-day advance deadline still described in older MFDS English guidance. Registration is normally valid for two years, with the statutory extension procedure. We supply the actual factory identity and supporting authorization documents early, and the importer verifies the approved registration and validity before filing. Requests for further information can lengthen the practical lead time, so the production and shipping schedule allows a buffer beyond the published service period. A requested private audit certificate has its own site and scope review.

Link source fields to Korean residue decisions

We propose beginning the residue file before harvest. Each planned broccoli source field should have an identifier, growing and harvest window, pesticide-use record, pre-harvest interval information where relevant, and incoming-lot connection to the processing run. We would review those records against the importer's target panel before assigning the field to the Korean order. A post-processing COA is still needed where agreed, but it cannot reconstruct a missing field identity or show what was sprayed if source lots were mixed without a documented boundary.

The Korean limit must be selected for the exact pesticide and commodity. The MFDS pesticide MRL portal classifies broccoli in the flowerhead-brassica group and says an individual commodity MRL overrides a group MRL. Its current rule applies 0.01 mg/kg when Annex 4 sets no pesticide/commodity limit. This is not a blanket claim that every tested active has a 0.01 limit: the importer must search the actual row first. A detected active may also have a legal residue definition covering parent and specified metabolites. We would have the Korean importer and laboratory agree the analyte names and legal definition before choosing a method or reporting "pass." The report should show measured compounds, any summing rule, units, recovery and limit of quantitation, not only a green check mark.

Broccoli source field and harvest traceability.

For perspective, the MFDS import-rejection portal records a fresh broccoli case in which pyraclostrobin was 0.66 mg/kg against a stated 0.05 mg/kg criterion. The example demonstrates an actual enforcement decision and the scale of an exceedance, but it must not be copied into this blanched frozen SKU as its current legal limit. We would use the present MRL portal and Food Code to lock the governing row for this product, then sample the actual source/production lot. Blanching and freezing are not presumed residue-removal processes. If a source field exceeds a Korean limit, changing the destination may be a lawful commercial option only after checking that destination's own rules; dilution into compliant material is not an acceptable assumed fix.

Our proposed test release record would identify the field or pooled source boundary, harvest and production lot, laboratory, method, sample date, analytes, individual results and Korean limit version used by the buyer. The buyer can add a safety margin or wider panel by contract; we should state that margin separately from law. If the laboratory LOQ is higher than a relevant default or specific MRL, a "not detected" result may be uninformative. We would hold the lot and seek a suitable method or a verified source alternative rather than call a weak test negative.

Trace blanching, cooling and IQF

Our proposed production dossier follows the lot from raw broccoli intake through trimming, cutting, washing, blanching, cooling, draining, IQF freezing and bulk packing. At intake we would link supplier and field IDs to gross weight, visible condition and a rejection decision for decay, insect damage, excessive soil or unsuitable maturity. The processing record should show when source lots enter and leave the run, who approves line changeovers and how product is segregated if a quality or test hold is issued. A carton code should lead back to those records without relying on memory.

Blanching has a legitimate quality purpose: controlling enzymes that can degrade colour and texture during storage. Its actual temperature, residence time, load, product geometry and cooling conditions matter to the resulting food. We should not invent a standard blanch setting or imply a validated kill step for this facility. The Korean plant-quarantine question is especially sensitive. The QIA 2026 field rulebook, page 330 describes cut plant material reaching 90°C, then being steamed or blanched for at least five minutes, followed by freezing to −17.8°C or colder. The Korean wording is "식물체를 절단하여 온도가 90℃에 도달한 시점에서 5분 이상 찌거나 데친 후 영하 17.8℃ 이하로 냉동처리한 과실류 또는 채소류"; an older QIA PDF explicitly says plant-body temperature. The current e-book viewer is unavailable at this check, so the importer must obtain QIA's current operative text and an item-specific decision. Ordinary blanching is not evidence that this narrow exclusion is met.

Broccoli florets on clean processing conveyor.

For any claimed QIA treatment, we would need actual batch measurement location, product temperature, dwell and post-freezing evidence for the specific lot. A hot-water or steam set point cannot stand in for the product temperature. Where that evidence is absent or QIA says quarantine applies, we would plan the required plant-health route instead of adding a certificate claim after loading. This is a conditional import decision, not a standard GreenLand process declaration. The buyer's quality blanch target, a food-safety validation target and a plant-quarantine exclusion target can be different; the order record should never collapse them into one unexplained number.

After heat treatment, cooling, water quality, drainage, transfer and freezing need hygienic and time/temperature control so cooked surfaces are not recontaminated before packing. The IQF objective is free-flowing florets within the agreed cluster tolerance, not an unsupported promise that every piece is identical. We would record process parameters, product checks, packing time, freezer/storage condition and deviations available for the actual order. Product temperature at packing and warehouse release should be measured by an agreed method. A freezer display alone records equipment air conditions, not every case core.

Use the cooking-required frozen-food category

The Korean importer and testing laboratory should confirm the exact Food Code classification before accepting a microbiology panel. This order is cooking-required and has not been represented as pasteurized or sterilized. The importer must first determine whether blanched broccoli is classified under this packaged processed frozen-food chapter or a more specific agricultural/processed-vegetable rule. If the non-pasteurized frozen-food branch applies, the MFDS frozen-food standard, the cooking-required non-pasteurized category lists aerobic plate count n=5, c=2, m=1,000,000, M=5,000,000 and E. coli n=5, c=2, m=0, M=10. For a pasteurized (살균제품) cooking-required frozen food, the same code instead gives APC n=5, c=2, m=100,000, M=500,000 and coliform n=5, c=2, m=10, M=100; the E. coli line excludes pasteurized products. The code defines pasteurization as product centre ≥63°C for 30 minutes or equivalent efficacy. MFDS distinguishes that from sterilization (멸균), which kills vegetative cells and spores. Our ordinary quality blanching description proves neither branch. These are five-unit acceptance schemes. The buyer and lab should identify the sample-unit basis, method, current code version and how results between m and M are counted. A supplier report that tests one convenient handful and prints the five-unit criteria is not equivalent evidence.

We would propose hygiene controls centered on the transition from blanch/cool to packing: cleaning and sanitation records, water controls, employee handling, equipment-contact surfaces, rework decisions and response to environmental findings. The buyer's own hazard assessment may require additional pathogen testing or a stricter indicator target, particularly if the florets enter a product with limited later heat treatment. Those requirements must be negotiated and documented; the existence of a cooking instruction does not excuse poor post-process hygiene. Conversely, a ready-to-eat category limit should not be applied to this SKU merely because it was blanched.

Frozen broccoli laboratory sample.

Our hold-and-release plan would identify the physical cartons covered by each sample and the time the laboratory result is due. Product should remain identified and unavailable for shipment while a contractual test is pending. If a result fails, we would review sampling chain, neighboring production lots, sanitation records and product disposition with the importer. Re-testing can clarify scope only under an agreed protocol; a second negative cannot automatically cancel a valid initial failure. Corrective action should say what changed on the line or source selection and what verification evidence proves the change effective before the next lot.

Microbiology is only one release layer. A lot can pass APC and E. coli but fail an MRL, packaging or foreign-body requirement. Equally, a few broken crowns can fail a visual purchase grade without making the food illegal. We would retain separate legal and commercial acceptance columns and name who may approve each exception. This keeps an MFDS inspection result, our proposed in-process check and the buyer's receiving decision from being treated as interchangeable certificates.

Separate appearance grading from foreign-body checks

The buyer should approve a representative cooked and frozen sample for colour, tenderness, crown diameter and stem proportion before the production target is locked. Our proposed visual-grading record would define yellowing, dark spots, decay, insect damage, leaves, excessive stalk and broken pieces. A single hero photo of selected florets is not a measurable shipment standard. We would use a stated sample size and lot distribution so a carton near the top of the pallet is not assumed to represent the entire run. Photos can document an agreed defect category, but a buyer's signed tolerance table and count/weight method are the actual acceptance tool.

Optical sorting, where used on the actual line, can separate material by visible colour and shape. It can support a greener, more uniform appearance and reject some visibly defective pieces. It cannot certify absence of invisible residues, pathogens or every foreign material. An X-ray system, where installed and validated for the actual package/product, detects density contrast and is aimed at specified foreign-body hazards. It is not simply a superior colour sorter. Metal detection addresses a different physical mechanism again. We would ask the processing site which controls are genuinely installed, the test-piece or challenge basis, sensitivity for the actual product, rejection verification and documented response to a failed check. We should never list hardware or detection thresholds that we have not verified for the selected facility.

Broccoli floret appearance grading.

The order evidence could include incoming raw-material inspection, trimming/sorting checks, foreign-body-control records, pack seal checks and finished-carton examination. The buyer should specify whether acceptance is per kilogram, carton, sample unit or full lot and how mixed defects are counted. Ice glazing, clumps or frost may be a cold-chain sign, not simply a cosmetic floret defect. We would record the location and time a defect is found so quality staff can distinguish field variation, processing loss, packing damage and transit abuse.

At shipment, we would preserve a sample or photo record that matches the production lot rather than selecting the best tray for sales. At destination, the buyer samples the agreed range of cartons while keeping stock frozen and identified. If a quality miss appears, both sides should compare the same definition and sampling basis before agreeing credit, replacement or a revised grade. Food-safety or legal signals remain on hold while the importer makes the regulatory decision; they should not be converted into a price discussion.

Approve a lot-level Korean release matrix

The practical release file starts with the purchaser's approved specification and the importer's legal decisions. It should state which check is mandatory for the shipment, which is conditional on MFDS/QIA classification and which is a negotiated buyer quality term. We would not mark every possible certificate as "required" or imply that our factory possesses one. The source field, manufacturing facility, carton label and commercial invoice must describe the same lot and product.

The importer should use the MFDS food-import declaration service for its filing. It permits submission as early as five days before expected arrival and shows no application fee; review periods differ by inspection track. Customs uses KCS UNI-PASS separately. We can assemble the product and shipment facts, but the Korean importer and broker approve and send official filings. A local COA cannot stand in for MFDS confirmation, and a customs message does not lift an unresolved plant-quarantine or internal QA hold.

GreenLand broccoli florets for lot release comparison.

For each row, the release decision needs a named owner and a date. If an MRL source changes during production, the importer should recheck the legal value before the laboratory report is judged. If the buyer changes from cooked foodservice use to uncooked garnish, the Food Code, pathogen and label review starts again. If a new factory makes the lot, its MFDS registration identity and our process assumptions must be revisited. Change control is less costly before cartons are printed and the container booked.

Ship cold and close the first-order learning loop

We propose an outbound inspection of sealed inner liners, sturdy cartons, legible lot marks, correct case count, clean pallets and a refrigerated container that is clean, dry and free from odour. The purchase specification should set the frozen delivery target and an agreed temperature-measurement method. We would link reefer ID, seal number, loading time, pallet/carton range and available logger position to the order. A set point is a control instruction, not proof that every floret remained at that exact temperature. The Korean buyer should specify what logger excursion, probe result or carton condition triggers a receiving hold.

At the port and bonded cold store, the importer coordinates MFDS inspection, any QIA decision and customs clearance while preserving frozen custody. The container or selected cartons may be opened under an inspection instruction; the warehouse should record who opened them, how they were protected and which lot/carton units were sampled. We would reserve a response contact for source records and avoid a promise that an ordinary declaration will always be completed in a fixed number of days. The Government24 MFDS service lists different administrative targets for document, field and close laboratory tracks, but terminal, sampling, QIA and customs events can extend the practical release.

Sealed frozen broccoli cartons in reefer.

Receiving quality staff should compare container seal, carton identity, physical count and temperature history with the approved lot dossier. They can open a representative sample to assess size, colour, stem and breakage under the contracted method. Frozen-product appearance should be judged consistently: condensation after an uncontrolled thaw can create misleading surface ice, while a cold-store sample can hide a temperature excursion that affected another pallet. A documented sampling location and timestamp allow both sides to reconstruct a complaint. The lot remains unavailable for use until the importer has the MFDS confirmation, QIA disposition if applicable, customs acceptance and its own receiving approval.

If the first shipment misses an agreed visual or packaging target, we would identify whether the cause lies in field maturity, trimming, sorting, packing compression or transport. We can then propose a trial with a changed size band, carton pattern or grade allocation and show the buyer its verification record. If the issue is an official residue or microbiological failure, the Korean importer controls the authority response and disposition; we would isolate related lots and investigate source and process. Re-export, destruction or a recall cannot be replaced by a supplier credit note. The next purchase order should carry the corrective action as a measurable new acceptance condition.

GreenLand-food can discuss a factory-direct wholesale proposal for this exact China-origin broccoli form, with quantity, bulk pack, delivery window and requested evidence. We would ask the Korean buyer to share its cooking application, import-business details, destination warehouse and legal/quality panel before a sample is approved. This keeps the offer specific enough to review. Any facility certificate, thermal validation or QIA exemption must be demonstrated for the actual factory and lot; it is not assumed from our marketing material.