D2, D3 and D4 wood glue — which should you choose? (2026)

A joiner chooses D3 glue for a garden table. Two winters later he is sitting with a table that slides apart every time someone lifts a glass. The joints have not broken — but the glue has softened and crept under the year's moisture and freeze cycles. He has to start over. Had he chosen D4, the table would still be standing in 25 years.

D-classification according to EN 204:2016 is the most important specification in wood bonding — and the one most joiners misinterpret. All PVA-based wood glues in Europe are classified as D1, D2, D3 or D4 according to how much moisture the joint can tolerate. The difference between the classes is not just "a little more water" — it is an entirely different polymer chemistry, entirely different test requirements and often a tenfold longer service life in an outdoor environment.

We distribute Titebond in Sweden and also have our own line of wood glue (D2 Inne, D2 Vinter, D3 Vattenfast) in various sizes from 750 ml up to a 100 L professional pack. We have seen hundreds of complaints where the root of the problem was precisely the wrong D-class. This guide is technically deep but practically angled: which class you need, why the tests are the way they are, and how to combine D-classes smartly in the same project. For a deep dive into the Titebond range, see our complete Titebond guide.

EN 204 — the standard in detail

EN 204 is the European harmonised standard that classifies thermoplastic wood adhesives for non-load-bearing applications by moisture resistance. It has existed since 1975, was thoroughly updated in 2001 and in its latest version EN 204:2016 — which is currently in force in the EU. In parallel, the actual test method is set out in EN 205, which describes exactly how beech test strips are glued, conditioned and pulled apart in a shear strength rig.

The distinction matters: EN 204 defines the requirements (which classes exist and what they must achieve), EN 205 describes how the test is carried out. A glue manufacturer's TDS that says "D3 according to EN 204" should always be backed by documentation according to EN 205.

Important to understand: EN 204 applies to non-load-bearing applications. For load-bearing structural bonding in glulam beams, EN 301 applies instead (for PRF, MUF and polyurethane structural adhesives). A D4 Titebond is not a structural adhesive — it is for furniture, windows, doors and joinery.

Alternative standards you may come across:

• ANSI/HPVA Type I and Type II — American standard. Type I is comparable to D4, Type II roughly D3. Titebond III is classified as Type I.
• ISO 17189 — international standard for thermoplastic wood adhesives, closely overlapping with EN 204 but not identical.
• JIS K 6806 — Japanese equivalent.

Boiling water is not realistic for a kitchen table — but it is a controlled, reproducible **simulation of the worst conceivable moisture and heat stress**. Freeze cycles would require different tests. The philosophy of EN 204 is: if a joint can withstand 100°C water, it can withstand any reasonable civilian outdoor environment.

The four load groups — the test sequences in detail

Each D-class has a specific test sequence according to EN 204:2016. The sequence numbers are formally called 1 to 5 and are combined according to the table in the standard. Here is what it looks like in practice:

D1 — permanently dry indoor environment

Conditioning: 7 days at 23°C and 50 % RH. Shear test thereafter.

Requirement: ≥ 10 N/mm² shear strength.

Use: Bookshelves, wardrobes, bedroom furniture — anything that never sees water or condensation.

D2 — occasional moisture, short-term exposure

Sequence: D1 conditioning (7 days dry) → 3 hours immersion in 23°C water → 7 days reconditioning dry → test.

Requirement: ≥ 8 N/mm² on the reconditioned test strip.

Use: Indoor joinery with a risk of condensation or spills — cabinets, chairs, tables, entrance-hall furniture.

D3 — variable moisture, sheltered outdoor environment

D3 is tested in two sequences:

1. D1 conditioning → 4 days in 23°C cold water → test directly on the wet test strip → requirement ≥ 2 N/mm².
2. The same sequence plus 7 days reconditioning → test on dry test strip → requirement ≥ 8 N/mm².

Combined: D3 must therefore handle both being wet and recovering. Typical values for a modern D3 glue lie around 11–13 N/mm² after reconditioning — well above the minimum.

Use: Kitchens, bathrooms, utility rooms, windows, doors, furniture under a roof outdoors, kitchen worktops.

D4 — full outdoor exposure

Sequence: D1 conditioning → 6 hours in boiling water (100°C) → 2 hours in 23°C water → test on the wet test strip.

Requirement: ≥ 4 N/mm² on the wet test strip.

This is the test that separates D3 from D4. Subjecting a glue to boiling water for 6 hours breaks down all standard PVA joints — only glues with modified chemistry (crosslinkers, polyurethane, resorcinol) can handle it. The test is a stress pulse, not a real-world use case.

Use: Garden tables, outdoor stairs, external doors, window frames freely exposed, above-deck boat fittings, cutting boards (together with FDA approval).

<CompareTable caption="EN 204:2016 in a single table" columns={[ { key: "klass", label: "Class", highlight: true }, { key: "test", label: "Test sequence" }, { key: "min", label: "Min. strength" }, { key: "bruk", label: "Use" } ]} rows={[ { klass: "D1", test: "7 days 23°C / 50 % RH", min: "≥ 10 N/mm²", bruk: "Permanently dry indoors" }, { klass: "D2", test: "+ 3 h water + 7 d reconditioning", min: "≥ 8 N/mm²", bruk: "Occasional moisture, indoors" }, { klass: "D3", test: "+ 4 d cold water", min: "≥ 2 N/mm² wet / ≥ 8 N/mm² dry", bruk: "Kitchens, bathrooms, outdoors under roof" }, { klass: "D4", test: "+ 6 h boiling + 2 h cold", min: "≥ 4 N/mm² wet", bruk: "Free outdoor exposure" } ]} />

The chemical difference — why D4 is different

All PVA wood glues are not the same polymer. The classes differ at the polymer level:

D1 and D2 are typically built on pure polyvinyl acetate homopolymer (PVAc). Curing is purely physical — the water evaporates, the polymer chains pack together, the joint becomes strong but remains water-soluble even after curing.

D3 uses PVA + VAE copolymer (vinyl acetate-ethylene). The ethylene unit makes the polymer softer and more hydrophobic — it lets less water in. More modern D3 glues also contain small amounts of external hardeners (isocyanate or melamine) that partially link the polymer chains, but without being fully covalently networked.

D4 is the big chemical leap. A typical D4 PVA glue contains:

1. PVA homopolymer as the main film-forming component.
2. VAE copolymer for hydrophobicity.
3. Functional monomers such as N-methylolacrylamide (NMA) built into the polymer chain.
4. Crosslinker catalyst — most commonly aluminium chloride in combination with a dicarboxylic acid (oxalic acid), although zinc-based systems also occur. Lewis-acid catalysis activates the NMA groups and creates covalent crosslinks between polymer chains as the glue cures.

It is the crosslinking — the actual chemical bonds — that means a D4 joint does not dissolve in boiling water. A D3 joint is held together by intermolecular forces (hydrogen bonds, van der Waals). A D4 joint additionally has a covalent network.

Titebond III Ultimate is Franklin International's proprietary D4 formulation — the exact chemistry is patent-protected, but the underlying architecture follows the pattern above. Titebond Polyurethane is an entirely different technology: a polyurethane prepolymer that reacts with moisture in the wood and the air and forms an expanding foam. Polyurethane is D4 but requires damp wood (8–15 % moisture content) to cure properly.

What the numbers mean — 4 N/mm² in practice

The EN 204 requirements are expressed in N/mm² shear strength. What does that really amount to?

• 1 N/mm² = 1 MPa = ~10 kg per cm²
• The D4 requirement after boiling water: 4 N/mm² = 400 N/cm² = ~40 kg per cm²
• A glue joint of 30 × 30 cm (e.g. tabletop to underframe): 900 cm² × 40 kg = 36 tonnes of remaining load-bearing capacity after the worst-case test.

Compare this with the inherent shear strength of the wood itself: oak has about 10–13 N/mm² shear strength along the grain, pine about 7–9 N/mm². A healthy Titebond III dry value is 25–30 N/mm² (from Franklin's own TDS) — three times the wood's inherent strength. Even after boiling-water stress D4 is at 4–8 N/mm², which is ~50 % of oak's shear strength.

In practice: there is no furniture load where a D4 joint is the limiting factor. The joint never fails first — the wood fails beside the bond line. That is the rule in all correctly executed PVA bonding.

When you pressure-test a glue joint, the failure should run through the wood, **not** through the glue line. If the glue joint is the weak point, you have either (a) the wrong glue class, (b) too thin a glue film, (c) insufficient clamping pressure, or (d) skipped conditioning. The EN 205 test only counts the **wood-failure percentage** as a secondary measure — the primary measure is N/mm².

The decision tree — how to choose the right class

Question 1: Where is the project going to stand?

• Permanently indoors, dry (living room, bedroom, office) → D2
• Indoors with moisture (kitchen, bathroom, utility room, hall by the front door) → D3
• Outdoors under a roof (patio, covered deck, carport) → D3 at least, preferably D4
• Outdoors with free exposure → D4 without exception
• Direct food contact (cutting boards, salad servers) → D4 + FDA approved (Titebond III)

Question 2: How critical is the joint?

• Decorative moulding that can be replaced → the previous class is enough
• Load-bearing furniture construction meant to last 20+ years → choose one class above the minimum
• Safety-critical (child's staircase, heavy shelf above a bed) → D3 indoors, D4 outdoors, ALWAYS

Question 3: How much time do you need to assemble?

• Simple joint, quick assembly → any class
• Complex assembly with many glue surfaces (e.g. finger-jointed tabletop) → choose a glue with longer open time — often D4 or specially classed D3 (Titebond Extend)
• Cold conditions (< 15°C) → choose a glue with higher temperature tolerance — Titebond III works down to 7°C, Titebond Original requires 10°C+. For Swedish winter workshops there is also Gjöco's D2 Vinter formulation, which film-forms at lower temperatures.

Concrete examples — 10 projects and the right class

1. Bookshelf in the living room → D2. Titebond Original or Gjöco Trälim D2 Inne.
2. Kitchen cabinets and cabinet doors → D3. Steam from cooking and washing up requires more than D2.
3. Kitchen worktop in solid wood → D3 + oil. Titebond II Premium or Gjöco Trälim D3 Vattenfast.
4. Utility-room cabinets → D3 with margin.
5. Window trim under a canopy → D3. Even sheltered air humidity is too much for D2.
6. Garden table → D4. Titebond III — non-negotiable.
7. Outdoor shower → D4 + pressure-treated wood and epoxy on the joints that sit below the water line.
8. Birdhouse or park benches → D4. Extreme freeze-cycle stress.
9. Cutting board or salad servers → D4 + FDA. Titebond III is the only wood glue certified to meet both.
10. Boat — fittings above deck → D4. Boat deck or below the waterline → epoxy or resorcinol, not PVA.

Bonus examples:

• A joinery course where the pupils build, rework and build again → D2. The short service life does not justify D3.
• Guitar bonding → D2 hide glue or fish glue traditionally. PVA will do but is irreversible — difficult to repair during rebuilds. The luthier standard is still hide glue.
• Church pew with extreme temperature swings → D4. An unheated church can swing from -15°C to +25°C over the year. D3 creeps under freeze cycles.

For **permanently submerged wood** (boat hulls below the waterline, foundations in water, lock elements) even D4 is not enough. Choose **epoxy** (West System, Sika Biresin) or **resorcinol-phenol glue** (historically used for boatbuilding) instead. PVA polymers always swell over time under permanent moisture exposure — only the rate differs.

How to test a glue class yourself — a simple workshop method

Amateur testing does not replace EN 205, but it gives you a good gut feel for how your glues behave. Do this:

<HowToSteps steps={[ { name: "Glue four test pieces", text: "Take eight clean beech pieces 20 × 150 × 5 mm. Glue them together in pairs (four pairs) with different glues — same clamping pressure, same technique, one of each class you are testing. Label the pairs." }, { name: "Dry 7 days at room temperature", text: "Leave the pairs at 23°C and roughly 50 % RH for a full week. This is the EN 205 D1 conditioning and a fair starting point." }, { name: "Immerse for 4 days in cold water", text: "Place one test pair in a tub of water for four days. This is the D3 sequence. Leave another pair for 6 hours in boiling water (D4 sequence)." }, { name: "Bench test with a punch and weight", text: "Clamp the piece in a bench vice, set a punch against the glue joint and load with weight until the joint breaks. Document the weight. A D3 joint after 4 days of water should hold at least 10 kg/cm²; a D4 joint after boiling at least 40 kg/cm². Far from lab precision, but you see the ranking straight away." } ]} />

Comparison — Titebond and Gjöco

Two wood-glue ranges that cover the entire D-span for Swedish joinery: the American premium brand Titebond (Franklin International, founded 1935) and the Norwegian Gjöco, which we distribute in Sweden.

<CompareTable caption="D-classes in Titebond and Gjöco" columns={[ { key: "brand", label: "Manufacturer", highlight: true }, { key: "d2", label: "D2" }, { key: "d3", label: "D3" }, { key: "d4", label: "D4" }, { key: "ot", label: "Open time" } ]} rows={[ { brand: "Titebond (USA)", d2: "Original", d3: "II Premium", d4: "III Ultimate", ot: "5–10 min" }, { brand: "Gjöco (NO)", d2: "Trälim D2 Inne", d3: "Trälim D3 Vattenfast", d4: "— (use Titebond III)", ot: "5–10 min" } ]} />

Gjöco is a Norwegian manufacturer of adhesives and surface finishes that we distribute for the Nordic market — cost-effective domestic bonding for D2 and D3 projects without compromising on D-classification. D2 Inne is available in 750 ml, 5 L, 15 L and 100 L (a professional pack for joineries pushing large volumes), and a dedicated D2 Vinter formulation (5 L) for workshops with lower temperatures. D3 Vattenfast comes in 750 ml and 5 L.

Titebond owns the D4 segment — Titebond III Ultimate is the glue we recommend for all outdoor exposure and is the only wood glue that is FDA-certified for indirect food contact (cutting boards, salad servers).

Price (guideline 2026, per litre):

• D2: 180–220 SEK
• D3: 240–290 SEK
• D4 PVA: 320–380 SEK
• D4 polyurethane: 400–500 SEK (and it expands — goes further)

Open time: Titebond III has about 8–10 min open time, Titebond II about 5 min. Gjöco D3 Vattenfast is around 5–10 min. At colder temperatures open time increases for all glues (rule of thumb: it doubles for every 10°C lower).

Five common misconceptions

1. "D3 is enough for anything outdoors as long as it is under a roof"

Wrong. Under a roof means protection from direct rain, but not from temperature and humidity swings that creep between 30 and 95 % RH over the year. A D3 glue handles rain for a month, not a decade of Swedish weather. Choose D4 if the wood is outdoors — regardless of the roof.

2. "D4 is just more expensive D3"

Wrong. D4 is a different chemistry. Titebond III has modified VAE copolymers and aluminium-chloride-catalysed crosslinkers that do not exist in D3. The difference is not marketing — it is documented EN 205 testing with a covalent network.

3. "PVA glue never becomes waterproof"

Partly right. No PVA becomes permanently waterproof — even D4 swells if it sits in water for months. But D4 handles normal outdoor exposure (rain, snow, freeze cycles) for 15–30 years. For permanent underwater use: epoxy or resorcinol.

4. "The highest class is always the safest — always go D4"

Wrong. D4 has shorter open time (usually), a darker glue joint on light wood, and a higher price. On indoor furniture where D2 is enough, D4 is wasteful and may also hurt the finish. Choose the right class, not always the highest.

5. "The D-class says everything about the glue"

Wrong. The D-class only tells you about moisture resistance according to EN 204. Other important properties — clamp time, open time, viscosity, film formation, FDA approval, creep resistance, UV tolerance — are independent of D-class. A cheap D3 and a premium D3 differ in all of these apart from moisture resistance. Read the full technical data sheet.

Why you should not glue D2 projects with D4

D4 glue has **shorter open time** (typically 8–10 min for Titebond III vs 20–25 for total clamping), a **darker glue joint** (Titebond III is brownish vs Original's yellow), and **higher cost**. On an untreated pine carcass where the glue joint will be visible, D4's dark line becomes an aesthetic minus. On a complex assembly where you need 15 minutes to get everything in place, D4 is a straight defeat. Use the right class where it is needed — not always the highest.

Summary — one bottle of each

Most workshops do perfectly well with three bottles on the glue shelf:

• D2 for indoor furniture — Titebond Original or Gjöco Trälim D2 Inne.
• D3 for the kitchen and damp areas — Titebond II Premium or Gjöco Trälim D3 Vattenfast.
• D4 for outdoors and cutting boards — Titebond III Ultimate.

Further reading: our complete Titebond guide goes through the entire product range. For a head-to-head comparison of Titebond III and the polyurethane alternative, see Titebond III vs Danalim Polyurethane Glue.

Sources

1. EN 204:2016 — Classification of thermoplastic wood adhesives for non-structural applications. CEN / SIS. standards.iteh.ai/catalog/standards/cen/en-204-2016
2. EN 205:2016 — Adhesives. Wood adhesives for non-structural applications. Determination of tensile shear strength of lap joints. Test method for EN 204.
3. Franklin International — Titebond III Ultimate Technical Data Sheet. titebond.com/product/glues
4. Industrieverband Klebstoffe — Interpretation of DIN EN 204/205. Düsseldorf, October 2017. klebstoffe.com — Interpretation_DIN_EN_204-205_EN.pdf
5. Vukov, Popović et al. (2016). Innovation in poly(vinyl acetate) water resistant D3 glues used in wood industry. International Journal of Adhesion and Adhesives. sciencedirect.com/science/article/abs/pii/S0143749616301269
6. US Patent 5091458A — Wood glue. Franklin International. Describes the aluminium-chloride/dicarboxylic-acid catalyst for PVA crosslinking. patents.google.com/patent/US5091458A
7. Woodbond — Tests for D3 Classification based on EN 205. Test-sequence documentation. woodbond.gr/D3-D4-category-Pvac-EN.pdf