Back to Research Hub
Implementation 11 min read

Sweden’s Siptex Facility: Near-Infrared (NIR) Sorting Driven by Woven Label Data Alignment

Exploring the inner workings of Sweden's Siptex center, the world's first automated industrial-scale NIR sorting facility.

D
DPPtex Research Team Published on July 5, 2026

The global textile recycling industry is caught in a paradox. While consumer demand for circular fashion surges, the technical reality of sorting post-consumer waste remains the primary bottleneck to scalable textile-to-textile recycling. Mechanical recycling, which shreds fabrics into fibers, is limited by contamination and fiber length degradation. Chemical recycling, which dissolves polymers, demands near-perfect feedstock purity—a single polyester button in a cotton load can ruin an entire batch. This is where Sweden’s Siptex facility, operated by Sysav, has emerged as a critical infrastructure node. Siptex is the world’s first large-scale plant capable of sorting textiles by fiber composition at industrial speeds using Near-Infrared (NIR) spectroscopy. However, the facility’s true breakthrough is not merely the hardware; it is the data alignment layer. By requiring manufacturers to sew in specialized NIR-active woven labels that reflect distinct light spectrums corresponding to exact fabric compositions, Siptex bridges the gap between the broad, high-volume search term “Textile Recycling” and the granular, compliance-driven reality of Digital Product Passports (DPPs). This article dissects the technical, regulatory, and supply chain architecture that makes this possible, offering a blueprint for importers and exporters navigating the EU’s circular economy mandates.

The Regulatory Framework & Macroeconomic Landscape

The Siptex model is not an isolated innovation; it is a direct response to a cascade of European regulatory deadlines that are reshaping the global textile trade. The most immediate driver is the EU’s Ecodesign for Sustainable Products Regulation (ESPR), which will mandate Digital Product Passports for all textiles placed on the EU market by 2030. Specifically, the delegated acts for textiles, expected in 2025, will require that fiber composition data be machine-readable and verifiable at the point of sorting. France’s AGEC Law (Article 13) already requires extended producer responsibility (EPR) for textiles, with mandatory sorting and recycling targets that penalize non-compliant importers. Germany’s Supply Chain Due Diligence Act (LkSG) adds a layer of forced labor and environmental auditing, meaning that the data on a label must now trace back to the dye house and the cotton farm.

From a macroeconomic perspective, the European Commission estimates that less than 1% of textile waste is currently recycled fiber-to-fiber. The rest is downcycled (insulation, rags) or incinerated. The cost of this inefficiency is staggering: the European Environment Agency calculates that textile waste costs the EU economy €12 billion annually in lost material value. Siptex directly addresses this by producing 24 distinct output fractions, including pure cotton, polyester-cotton blends, and wool, which can command prices 3–5x higher than mixed waste. For Swedish and Nordic fashion brands partnering with Sysav, this feedstock is the only viable path to meeting their 2030 circularity targets without resorting to greenwashing. The compliance deadline is unforgiving: by 2028, the EU Waste Framework Directive will require member states to achieve separate collection of textiles, and by 2030, a minimum recycling rate of 55% for textile waste.

Deep Supply Chain Execution & Exporter Challenges

The technical elegance of Siptex’s NIR sorting masks a brutal reality for exporters in manufacturing hubs like Bangladesh, Vietnam, Sri Lanka, Turkey, and Brazil. The requirement for NIR-active woven labels—which must be sewn into every garment as a permanent, non-removable identifier—forces fundamental changes on the factory floor. The Bangladesh Garment Manufacturers and Exporters Association (BGMEA) has flagged this as a critical capacity gap: most factories currently use printed care labels that are chemically inert under NIR light. Retooling to produce woven labels with specific polymer additives (e.g., rare-earth oxides that reflect at 1200–1400 nm) requires new looms and quality control spectrometers, a capital investment of $50,000–$100,000 per factory.

Local constraints compound the challenge. In Vietnam, the VITAS association reports that 60% of textile factories face unstable energy grids, making the continuous operation of NIR label printers—which require precise temperature and humidity control—unreliable. In Sri Lanka, the Joint Apparel Association Forum (JAAF) has noted that informal labor practices in label sewing stations lead to high error rates (up to 15% mislabeling), which would cause Siptex’s NIR sensors to misclassify entire bales. Turkey’s ITHIB has responded by establishing a centralized label certification hub in Istanbul, where labels are pre-validated against a reference spectral library before shipment. Brazil’s ABRAPA is experimenting with embedding NIR-active threads directly into the seam tape, bypassing the label entirely.

Technologically, the exporter must now integrate a data pipeline that was previously nonexistent. The label’s NIR signature must be registered in a GS1 Digital Link QR code or an NFC chip that links to a DPP. This requires factories to install RFID printing stations (e.g., Zebra ZT600 series) and secure a stable internet connection to upload the label’s spectral fingerprint to Sysav’s API. The failure mode is stark: if the label’s NIR signature does not match the DPP’s declared composition, the garment is rejected at the Swedish border and returned at the exporter’s cost. This is not a theoretical risk; Sysav’s 2023 technical disclosure reported a 7.2% rejection rate for the first 500,000 garments sorted, primarily due to label misalignment.

Data Specifications & Testing Benchmarks

The following table maps the critical data fields required for Siptex compatibility, the corresponding test methods, and the validation roles for each stakeholder.

Data FieldDescriptionTest Method / StandardValidating Role
NIR Label Spectral SignatureUnique reflectance curve (900–1700 nm) of the woven label additiveISO 20473:2007 (Optics and photonics — Spectral bands)Exporter’s label supplier (certified by Sysav)
Fiber Composition (Mass %)Exact blend ratio (e.g., 65% Polyester, 35% Cotton)ISO 1833:2020 (Quantitative chemical analysis)Third-party lab (ISO 17025 accredited)
Fabric Weight (g/m²)Areal density of the base fabricISO 3801:1977 (Textiles — Woven fabrics — Determination of mass per unit length and mass per unit area)Factory QA
Dye & Finish ChemistryPresence of heavy metals, PFAS, or antimonyISO 4484-1:2023 (Textiles and textile products — Microplastics from textile sources)Chemical supplier (REACH compliance)
Garment Unique ID (UID)GS1-128 barcode or EPC UHF RFID tagGS1 General Specifications v24 / ISO 18000-6CExporter’s IT system
DPP Digital Twin URLHTTPS endpoint resolving to a Verifiable CredentialW3C DID Core 1.0 / EPCIS 2.0 JSON-LDImporter’s DPP platform (e.g., Trustrace, TextileGenesis)
Lifecycle Assessment (LCA)Cradle-to-gate carbon and water footprintISO 14040:2006 / ISO 14044:2006Brand’s sustainability team
Supplier Audit ReportForced labor and environmental complianceGerman LkSG / US UFLPA guidelinesAccredited auditor (e.g., SGS, Bureau Veritas)

Detailed Technical Architecture Block

The following ASCII flowchart illustrates the physical-digital scanning loop at Siptex, from label detection to DPP resolution.

+-------------------+       +-------------------+       +-------------------+
|   Exporter Factory|       |   Siptex Facility |       |   Sysav API Cloud |
|                   |       |   (Malmö, Sweden) |       |                   |
|  Sew NIR label +  |       |                   |       |                   |
|  Encode UID in QR |       |   Conveyor Belt   |       |   DPP Resolver    |
|  Upload DPP to    |       |   (3 m/s)         |       |   (GS1 Digital    |
|  Sysav API        |       |                   |       |    Link)          |
+--------+----------+       +--------+----------+       +--------+----------+
         |                           |                           |
         | 1. Garment arrives        |                           |
         |    with NIR label         |                           |
         |-------------------------->|                           |
         |                           | 2. NIR sensor scans      |
         |                           |    label (900-1700 nm)   |
         |                           |    -> Spectral match?    |
         |                           |    Yes/No                |
         |                           |                           |
         |                           | 3. If Yes: Read UID      |
         |                           |    from QR/RFID tag      |
         |                           |    -> Send UID to API    |
         |                           |-------------------------->|
         |                           |                           |
         |                           |                           | 4. API resolves UID
         |                           |                           |    to DPP JSON-LD
         |                           |                           |    -> Validate fiber
         |                           |                           |    composition matches
         |                           |                           |    NIR signature
         |                           |                           |
         |                           | 5. Return sorting        |
         |                           |    instruction (bin #)   |
         |                           |<--------------------------|
         |                           |                           |
         |                           | 6. Pneumatic diverter    |
         |                           |    routes garment to     |
         |                           |    correct bin (cotton,  |
         |                           |    poly-cotton, wool...) |
         |                           |                           |
         | 7. Sorting confirmation   |                           |
         |    logged to DPP          |                           |
         |<--------------------------|                           |

Below is a valid JSON-LD metadata payload representing the DPP data packet that must accompany each garment. This payload is resolved by the Sysav API in step 4 above.

{
  "@context": {
    "@vocab": "https://w3id.org/epcis/",
    "gs1": "https://gs1.org/voc/",
    "dpp": "https://example.org/dpp/v1/"
  },
  "@id": "urn:epc:id:sgtin:0614141.812345.6789",
  "@type": "EPCISDocument",
  "epcisBody": {
    "eventList": [
      {
        "@type": "ObjectEvent",
        "eventTime": "2025-03-15T10:30:00.000Z",
        "eventTimeZoneOffset": "+01:00",
        "action": "OBSERVE",
        "bizStep": "urn:epcglobal:cbv:bizstep:sorting",
        "disposition": "urn:epcglobal:cbv:disp:in_progress",
        "readPoint": {
          "id": "urn:epc:id:sgln:0614141.00001.0"
        },
        "bizLocation": {
          "id": "urn:epc:id:sgln:0614141.00001.0"
        },
        "epcList": [
          "urn:epc:id:sgtin:0614141.812345.6789"
        ],
        "bizTransactionList": [
          {
            "type": "urn:epcglobal:cbv:btt:po",
            "bizTransaction": "urn:epcglobal:cbv:bt:0614141:PO-2025-001"
          }
        ],
        "ilmd": {
          "dpp:fiberComposition": {
            "dpp:material": "Polyester",
            "dpp:percentage": 65,
            "dpp:material": "Cotton",
            "dpp:percentage": 35
          },
          "dpp:nirSignature": "0x7A3F9C1B2E4D5F8A",
          "dpp:labelType": "Woven_NIR_Active_Class_III",
          "dpp:spectralRange": "900-1700nm",
          "dpp:recyclingMethod": "Chemical",
          "dpp:eprScheme": "Sysav_SE_2025",
          "dpp:lcaReference": "https://api.sysav.se/dpp/lca/0614141.812345.6789",
          "gs1:netWeight": {
            "gs1:value": 0.250,
            "gs1:unitCode": "KGM"
          }
        }
      }
    ]
  }
}

Actionable Compliance Checklist

[!IMPORTANT] Mandatory Steps for Exporters and Importers to Achieve Siptex NIR Sorting Compatibility

  1. Label Procurement & Certification

    • Source woven labels from a Sysav-approved supplier that provides a Certificate of Spectral Conformance (CoSC) for each batch.
    • Ensure the label additive (e.g., lanthanide-doped polymer) is REACH and OEKO-TEX ECO PASSPORT certified.

  2. Factory Floor Retooling

    • Install a NIR spectrometer (e.g., Thermo Scientific microPHAZIR) at the final quality control station to verify label spectral signature before shipment.
    • Train sewing operators on label placement (must be in the side seam, not the neck label, to avoid NIR shadowing).

  3. Data Pipeline Integration

    • Deploy a GS1 Digital Link QR code generator that encodes the garment UID and resolves to a DPP hosted on a GS1-compliant resolver (e.g., GS1 Sweden’s Verify platform).
    • Implement an API webhook to Sysav’s POST /v1/sorting/manifest endpoint, sending the UID and NIR signature for each SKU.

  4. Pre-Shipment Validation

    • Conduct a dry-run test: ship 100 sample garments to Sysav’s test line in Malmö. Request a sorting accuracy report. Target: <1% misclassification.
    • Obtain a Third-Party Audit Report (ISO 17025) for fiber composition using ISO 1833:2020.

  5. Importers (Nordic Brands)

    • Register with Sysav’s EPR scheme and obtain a Producer Identification Number (PIN).
    • Ensure your DPP platform (e.g., Trustrace, TextileGenesis) supports the dpp:nirSignature field in the EPCIS 2.0 payload.
    • Negotiate a feedstock offtake agreement with Sysav for the sorted fractions (e.g., 100% cotton for mechanical recycling, poly-cotton for chemical recycling).

Strategic Conclusion

The Siptex facility represents a paradigm shift from waste management to resource intelligence. By forcing the alignment of physical labels with digital data, it creates a closed-loop verification system that solves the feedstock purity problem plaguing chemical recyclers. For exporters, the investment in NIR-active labels and DPP infrastructure is not optional; it is a condition of market access to the EU. For importers, the partnership with Sysav offers a defensible supply chain that can withstand regulatory audits under the ESPR and LkSG. The next frontier is the integration of this NIR sorting data into real-time carbon accounting, allowing brands to claim verified emission reductions from avoided incineration. As the EU moves toward mandatory DPPs, the Siptex model will likely become the de facto standard for textile sorting infrastructure across Europe, from the Netherlands’ Wieland Textiles to France’s CETIA. The message is clear: in the circular economy, the label is no longer just a care instruction—it is a machine-readable contract for the garment’s next life.

📚 Regulatory & Academic Bibliography