Establishing Non Target Screening Limits for Non Intentionally Added Substances in Complex Recycled Resin Matrices

Noise

Post-consumer polyolefin and polyethylene terephthalate flakes arrive at re-processing facilities carrying chemical residues from prior service, outdoor exposure, and heat history. During gas chromatography and liquid chromatography coupled with high-resolution mass spectrometry, this chemical background creates substantial baseline interference. Unresolved complex mixtures show up as broad baseline humps that bury low-abundance non-intentionally added substances migrating at toxicologically relevant levels.

In post-consumer recycled high-density polyethylene, broad oligomer distributions, oxidized antioxidant fragments, and sorbed volatile organics quickly saturate electron ionization and electrospray sources. This background raises detection limits and hides trace targets, rendering standard signal-to-noise calculations useless for compliance verification.

What gets extracted depends heavily on the solvent system chosen for sample prep. Dichloromethane pulls out broad lipophilic fractions ~ including polyethylene oligomers up to seven hundred Daltons that co-elute with primary antioxidant degradation products. Ninety-five percent ethanol mimics fatty food contact while extracting fewer total oligomers than chlorinated solvents, though it still pulls out polar degradation species and fragrance compounds.

Hexane isolates non-polar hydrocarbons well while leaving behind polar degradation compounds with low toxicological thresholds. Background signal thresholds are measured by comparing total ion chromatograms from blank solvent runs directly against polymer extracts under identical chromatographic conditions.

Chemical loads in recycled resins are unpredictable. When matrix components co-elute directly with migrating non-target analytes, baseline subtraction algorithms often break down. In liquid chromatography coupled with electrospray ionization mass spectrometry, matrix components alter analyte ionization directly via charge competition in the spray capillary.

Severe suppression can cut analyte response by over eighty percent, turning a compound present above toxicological concern into an invisible baseline dip. Conversely, signal enhancement overestimates concentration, prompting false rejections of compliant resin lots. Reliable non-target screening therefore depends on procedural blank subtraction, solvent blank monitoring, and matrix-matched spike corrections to establish true baselines.

Gas chromatography with flame ionization detection yields an unresolved complex mixture baseline signal exceeding fifty milligrams per kilogram in unwashed post-consumer polyolefin extractions.

Blank subtraction routines require strict algorithmic thresholds so real non-target peaks are not stripped out during data reduction. Data processing software checks peak areas in resin extract chromatograms against matching retention time windows in procedural blanks. Applying a fixed fold-change cutoff ~ such as flagging target peaks at five times the solvent blank intensity ~ removes routine laboratory contaminants like phthalates from solvent stock or siloxanes from GC septum bleed.

But in complex post-consumer resins, uneven background signals call for dynamic thresholding tied to local chromatographic noise rather than blanket peak-area cutoffs. Matrix interference remains the chief reason non-target identifications yield false negatives in food-contact recyclates.

Overlooking baseline chemical complexity during non-target screening allows non-compliant resin batches with uncharacterized toxic compounds to slip straight into food packaging converting lines.

Bench

High-resolution mass spectrometry hardware provides the raw analytical detail needed to untangle complex mixtures extracted from recycled polymers. Gas chromatography paired with quadrupole time-of-flight systems uses electron ionization at seventy electronvolts to produce reproducible fragmentation patterns for matching against commercial spectral libraries. On the liquid chromatography side, Orbitrap or time-of-flight instruments use electrospray or atmospheric pressure chemical ionization to capture intact protonated or deprotonated molecular ions.

Resolving power above forty thousand full width at half maximum at mass-to-charge ratio two hundred yields the mass accuracy needed to determine elemental compositions within a two parts per million window.

Spectral Deconvolution and Mass Resolution Capabilities

Co-eluting peaks in recycled resin extracts make automated spectral deconvolution essential for pulling out clean compound spectra. Deconvolution software tracks ion signals across chromatographic peaks, clustering ions that share identical retention times and peak shapes. This separates individual chemical species from the unresolved baseline hump, pulling co-eluting degradation products away from background polymer oligomers.

Accurate mass measurement of isotopic patterns confirms chlorine, sulfur, or bromine content, giving an immediate structural filter for halogenated flame retardants or legacy industrial additives.

Deconvolution isolates overlapping spectral signals, but mass accuracy alone guarantees very little. Without retention time confirmation and fragmentation matching, structural assignments remain tentative hypotheses.

Confidence Hierarchy in Non Target Structure Assignment

Structural identification follows the established five-level confidence framework developed for environmental and packaging non-target screening. Level one confirmation requires matching retention time, exact mass, and fragment spectra against an authentic reference standard run on the same instrument. Level two relies on exact mass, diagnostic fragments, and strong matches to spectral libraries or published literature.

Level three denotes a tentative structure where candidate isomers cannot be separated by fragmentation alone. Level four assigns a definitive molecular formula from isotopic patterns without proposing a specific structure, while Level five represents an exact mass feature isolated above baseline noise without an assigned formula.

Hardware Limits and Spectral Deconvolution Failure Modes

The analytical hardware selected directly determines which chemical classes can be detected. Workflows fail when target compounds fall outside instrument operating windows or when deconvolution routines misattribute baseline signals. The following list highlights common mechanical and analytical failure points observed during non-target screening of post-consumer resins.

• Incomplete thermal desorption leaves high-molecular-weight species above six hundred Daltons stranded inside gas chromatography injection liners, preventing quantification of heavy synthetic antioxidant degradation products.
• Electrospray ion suppression extinguishes ion formation for non-polar slip agents and oxidized wax fragments in liquid chromatography sources when co-eluting polyethylene oligomers saturate droplet charge capacity.
• Mass spectrum fragmentation over-clustering during hard electron ionization breaks fragile aliphatic molecules into non-specific hydrocarbon fragments, preventing automated library search routines from matching parent structures.
• Inadequate mass resolving power below twenty thousand full width at half maximum fails to separate doublet ion masses sharing identical nominal mass values, such as sulfur-containing antioxidant fragments versus pure hydrocarbon ions.
• Inappropriate chromatographic solvent selection introduces strong background ion clusters in electrospray ionization, masking mass features within the critical range of one hundred to three hundred Daltons.

A non-target screening report showing no detected peaks above one milligram per kilogram does not guarantee resin purity if the analytical method’s limit of detection sits well above that threshold for non-polar migrants.

Drift

Quantifying identified and unidentified non-target peaks without authentic reference standards is one of the hardest aspects of packaging safety verification. High-resolution mass spectrometry ionization efficiencies vary by up to three orders of magnitude across different chemical families. A fixed concentration of ten micrograms per kilogram produces drastically different peak areas for an aliphatic amine, a sterically hindered phenol antioxidant, and a neutral organophosphite degradation product.

Relying on a single internal standard like deuterated toluene or atropine to quantify every detected feature introduces unacceptable measurement error into safety assessments.

Relative response factor distributions define the mathematical boundary of semi-quantification uncertainty. To estimate unknown concentrations, analytical chemists divide an analyte’s detector response factor by that of an assigned internal standard surrogate. Measuring broad sets of model compounds spanning diverse functional groups, polarities, and molecular weights generates a statistical distribution of relative response factors.

Calculating relative response factors across eighty representative surrogate compounds establishes the tenth percentile relative response factor value as the conservative foundation for semi-quantification.

Derivation of the Analytical Screening Limit

Setting an analytical screening limit requires translating toxicological exposure thresholds into an instrument-specific signal cutoff concentration. The screening limit ensures that any chemical substance capable of migrating above its toxicological threshold generates a chromatographic peak area exceeding the integration trigger, even when it ionizes poorly.

The mathematical formulation for the analytical screening limit incorporates the toxicological threshold of concern, the exposure ratio of packaging surface area to food volume, the sample concentration factor achieved during preparation, and the lower-bound relative response factor percentile. The governing equation is expressed as:

ASL = (TTC CF) / (UF RRF_percentile)

Where ASL represents the analytical screening limit in milligrams per kilogram of food simulant or extract, TTC is the toxicological threshold of concern in milligrams per kilogram of food, CF is the food contact exposure factor (conventionally assuming six square decimeters of packaging contacts one kilogram of food), UF is an additional analytical safety uncertainty factor, and RRF_percentile is the lower tenth percentile relative response factor determined for the specific analytical platform.

Consider a practical calculation scenario for post-consumer recycled high-density polyethylene intended for fatty food packaging. The chosen toxicological threshold of concern corresponds to Cramer Class III compounds, which sets a maximum human exposure limit of 0.0015 milligrams per kilogram of food (1.5 micrograms per kilogram). The laboratory extracts ten grams of polymer into twenty milliliters of ninety-five percent ethanol, achieving a concentration factor where two grams of packaging material are represented per milliliter of extract.

Evaluating the liquid chromatography electrospray ionization platform across eighty calibration standards demonstrates a tenth percentile relative response factor of 0.12 relative to the internal standard d10-benzophenone. Inserting these values into the screening limit calculation yields:

ASL = (0.0015 mg/kg 1.0) / (1.5 0.12) = 0.00833 mg/kg

Converting this value to extract concentration based on the packaging-to-simulant contact ratio reveals that the analytical system must reliably detect and integrate peaks down to 0.00833 milligrams per liter (8.33 micrograms per liter). Peaks displaying areas corresponding to concentrations above 8.33 micrograms per liter must undergo mandatory identification or toxicological evaluation.

Analytical screening limits set without accounting for ionized response variability systematically underestimate structural alerts with weak ionization efficiencies.

Response factors fluctuate widely across compound classes. When analytical screening limits fail to incorporate lower-bound response factor adjustments, compounds exhibiting weak detector responses pass undetected even when present at levels ten times above safe toxicological thresholds.

Applying uniform internal standard response assumptions across non-target datasets remains a widespread but mathematically invalid shortcut in recycled resin screening protocols.

Tier

When full toxicological datasets are unavailable, evaluating non-intentionally added substances flagged during non-target screening relies on exposure-based risk thresholds. European Union legislation under Regulation EC 1935/2004 and Regulation EU 10/2011 mandates that all migrating substances undergo safety assessment, regardless of their origin. The Threshold of Toxicological Concern concept provides a standardized framework to evaluate chemical substances present at low concentrations without compound-specific toxicity data.

Chemical structures identified at confidence levels one through three are mapped to Cramer structural classes using decision-tree software such as Toxtree or the OECD QSAR Toolbox.

Cramer Classification Tiers and Exposure Cutoffs

Cramer Class I represents substances with simple chemical structures and efficient metabolic detoxification pathways, carrying a human exposure threshold of 1.8 milligrams per person per day, equivalent to thirty micrograms per kilogram of food assuming standard exposure parameters. Cramer Class II contains moderately complex structures that display less predictable toxicity, assigned a threshold of 0.54 milligrams per person per day (nine micrograms per kilogram of food). Cramer Class III encompasses complex chemical structures, aromatic amines, organophosphates, and structural alerts for toxicity, carrying a stringent threshold of 0.09 milligrams per person per day (1.5 micrograms per kilogram of food).

Compounds containing reactive functional groups known to cause direct DNA damage fall outside standard Cramer classes. Structural alerts for genotoxicity, such as alkylating agents, aromatic nitro groups, and simple aziridines, trigger the genotoxic threshold of toxicological concern set at 0.15 micrograms per person per day, corresponding to a concentration of 0.0025 micrograms per kilogram of food. Any unidentified peak or compound possessing a genotoxic structural alert that exceeds this threshold requires definitive identification, synthesis of authentic reference standards, and empirical mutagenicity testing using the Ames assay.

Technical dossiers lacking semi-quantitative response factor bounds face immediate rejection under European Union recycled plastic verification audits.

Structural Alert Evaluation Sequence

Evaluating unknown chromatographic peaks detected in post-consumer recycled polyolefins requires a systematic decision workflow. The following steps outline the toxicological evaluation procedure applied to screen unidentified and semi-identified peaks against exposure thresholds.

1. Deconvolution and Peak Quantification converts chromatographic raw peak areas into semi-quantitative concentration values using the tenth percentile relative response factor derived for the analytical system platform.
2. Genotoxic Structural Alert Screening checks candidate chemical structures identified at confidence levels two and three for electro-reactive functional groups using in silico toxicological prediction software.
3. Cramer Class Assignment maps non-genotoxic candidate structures to Cramer Class I, II, or III structural tiers using decision-tree rule sets.
4. Unidentified Feature Default Tiering assigns unidentified peaks exhibiting Level four or Level five confidence directly to the genotoxic threshold tier of 0.0025 micrograms per kilogram of food unless high-resolution isotope patterns rule out aromatic amines and alkylating functional groups.
5. Exposure Margin Assessment calculates the expected daily consumer intake based on packaging contact area, food shelf-life, fill volume, and food consumption habits.
6. Safety Factor Verification compares calculated exposure levels against assigned Cramer or genotoxic thresholds, triggering mandatory toxicological testing when exposure margins fall below one.

What Happens When Unknown Peaks Resist Toxicological Classification?

Unidentified mass features lacking clear molecular formula assignments present severe compliance bottlenecks under European food contact regulations. When high-resolution mass spectrometry fails to assign a candidate structure due to low signal intensity or absence of clear isotopic patterns, automated Cramer classification tools cannot operate. Under standard risk assessment protocols, unassigned features must default to the lowest applicable toxicological threshold, assuming potential genotoxicity at 0.0025 micrograms per kilogram of food.

This conservative default forces converters to perform extensive sample concentration or clean-up to prove absence, or to reject the underlying recycled resin batch entirely.

Alternative toxicological approaches attempt to group unassigned features into broad chemical families using fragmentation spectral patterns. High-resolution fragment spectra matching aliphatic hydrocarbon patterns permit assignment to the mineral oil saturated hydrocarbon toxicity tier, raising the tolerable migration threshold to 0.6 milligrams per kilogram of food. Similarly, identifying characteristic fragment ions associated with oxidized polyolefin oligomers allows resin evaluators to exit the genotoxic default threshold, applying oligomeric toxicity thresholds derived from synthetic polyethylene toxicology studies.

Determining whether spectral pattern grouping provides sufficient legal certainty under regulatory audit remains an open debate within compliance authorities.

How far can analytical chemistry reduce the fraction of unassigned features before post-consumer resin sorting costs exceed the market price of virgin polymer?

Batch

Post-consumer recycling streams exhibit high chemical heterogeneity compared to virgin polymer production. Sorting efficiency, seasonal collection variance, and contamination events cause chemical residue levels to fluctuate significantly between output lots. Single-point non-target screening on a single resin lot provides no statistical guarantee that subsequent production lots share identical NIAS profiles.

Establishing robust non-target screening limits requires integrated batch monitoring strategies that balance analytical testing frequencies with statistical process control.

Composite sampling strategies combine material samples across continuous extrusion runs to capture average chemical profiles. Individual flake samples taken at one-hour intervals across a twenty-four-hour extrusion cycle are blended prior to solvent extraction and mass spectrometry analysis. Composite sampling reduces total analytical testing burdens while maintaining sensitivity for continuous, low-level contaminants.

Composite sampling can dilute localized spike contamination events, diluting a hazardous contaminant present in one isolated pellet lot below the analytical screening limit.

Testing single batches creates blind spots. Statistical process control for recycled resin streams requires setting baseline action limits for overall chromatographic peak area counts and total ion counts. Tracking total non-target peak counts across consecutive production lots generates process capability charts.

A sudden spike in the total number of mass features detected above the noise floor indicates a breakdown in upstream feedstock sorting or decontamination washing efficiency, triggering an immediate hold on the affected lot prior to packaging conversion.

The following sequence outlines the statistical qualification protocol applied to incoming post-consumer recycled polyolefin resin batches before release into food-contact manufacturing streams.

1. Collect three independent five-hundred-gram resin samples from distinct spatial zones within each incoming gaylord container or bulk silo.
2. Homogenize collected samples through cryogenic grinding to achieve uniform particle size distribution below one millimeter.
3. Subject a ten-gram homogenized aliquot to rapid headspace gas chromatography coupled with flame ionization detection to quantify total volatile organic content against a five-milligram-per-kilogram toluene equivalent threshold.
4. Extract a secondary twenty-gram aliquot using ninety-five percent ethanol for ten days at sixty degrees Celsius to model worst-case fatty food contact migration.
5. Analyze extracts using high-resolution liquid and gas chromatography mass spectrometry, applying pre-established analytical screening limits adjusted for relative response factor uncertainty.
6. Compare generated mass feature tables against historical resin baseline databases using multivariate statistical analysis to identify outlying chemical signals.
7. Quarantine any batch exhibiting mass features exceeding the toxicological threshold of concern or displaying structural alerts for genotoxicity until definitive identification confirms safety.

Quality assurance contracts between recycled resin suppliers and packaging converters specify that any batch exhibiting uncharacterized non-target peaks exceeding 0.01 milligrams per kilogram food simulant equivalent shall trigger immediate lot rejection at the supplier expense.

Dossier

Declarations of Compliance issued for post-consumer recycled resins must reflect empirical non-target screening data rather than standardized boilerplate language. Under European Union Regulation EU 2022/1616 on recycled plastic materials intended to come into contact with food, recyclers and converters carry clear legal responsibility for the chemical safety of finished articles. Supporting technical dossiers must document the complete analytical workflow used to screen, identify, and toxicologically evaluate non-intentionally added substances.

Auditing authorities systematically reject technical dossiers that rely exclusively on targeted testing for regulated additives while ignoring non-target mass spectrometry profiles.

In technical compliance files, broad target lists frequently mask missing non-target data. A supplier declaration stating compliance with European Union Regulation EU 10/2011 Annex I positive lists provides zero legal protection against migrating un-listed degradation products or post-consumer contaminants. The technical file behind the declaration must explicitly state the limit of detection achieved for non-target screening, the relative response factor uncertainty factors applied during semi-quantification, and the toxicological decision trees used to evaluate unassigned mass features.

Without these specific parameters, the declaration remains an empty administrative statement.

Declarations of compliance that reference target testing alone leave importers fully liable for uncharacterized migrating species.

Customs authorities and food safety inspectorates enforce chemical compliance through targeted border holds and market surveillance sampling. Upon selecting a packaging shipment for audit, enforcement officers demand immediate access to the supporting technical file. If the file lacks explicit non-target screening data covering volatile, semi-volatile, and non-volatile migration fractions, authorities issue market withdrawal orders and border rejections.

Commercial liability for recalled product lines and stranded freight flows directly up the supply chain to the entity that signed the Declaration of Compliance without maintaining verified analytical back-up.

Supply chain contracts must allocate the financial risks associated with non-target chemical non-compliance. Sourcing agreements require clear technical specifications defining maximum permissible total non-target peak area counts, mandatory screening frequencies, and analytical methodology standards. Including explicit language that defines analytical screening limits based on tenth percentile relative response factor models ensures that both resin recycler and packaging converter operate under identical chemical safety standards.

Compliance demands verifiable analytical dossiers backed by traceable mass spectrometry data generated on the specific production lots loaded into export containers.

Managing chemical risk in complex recycled resin matrices requires aligning analytical bench capabilities with toxicological thresholds, statistical process controls, and legally defensible supply chain documentation.