The Glass Containers Passed Inspection. They Shattered the Moment They Went Into the Oven
Ceramic and Glass Thermal Shock Test Guide - Why Glass Containers and Ceramic Tableware That Passed Inspection Crack After Sale, Borosilicate vs Soda-Lime Glass, Water Quench, Oven and Microwave Thermal Shock Tests Compared, Reading Temperature Difference Criteria, Poor Annealing and Residual Stress, Ceramic Water Absorption and Glaze Crazing, Gaps in Factory Reports, Supplier Glass Substitution, Spec Sheet Items, Stage-by-Stage Checks and a Checklist

Hello, this is StartMade.

A Korean kitchenware brand imported 5,000 sets of "borosilicate heat-resistant glass" food storage containers from a glass factory in Shandong. The approval sample was clear with even wall thickness, and the pre-shipment inspection found no problems with appearance, dimensions or lid sealing. The product page said "Goes straight from freezer to oven."

About a month into sales, complaints started coming in. Customers said they had taken a container of side dishes out of the fridge, put it into a preheated oven, and heard a pop as the bottom split. When similar complaints kept arriving, the brand stopped selling. A lab measured the coefficient of thermal expansion on the remaining stock and got a value for ordinary soda-lime glass, not borosilicate. In a temperature difference test, the containers cracked at a 60°C difference. The factory had made only the approval samples in borosilicate and pressed production in soda-lime glass that costs about half as much. The two glasses are almost impossible to tell apart by eye.

Thermal shock failures are tricky because a room-temperature inspection never reveals them. Glass and ceramics break when a sudden temperature change makes the inside and outside expand and contract at different rates. Two pieces with the same shape and weight can withstand temperature differences several times apart, depending on the material, annealing and firing temperature. This article starts with why glass and ceramics crack from heat, then covers test types and how to read criteria, common supplier tricks, and how to build thermal shock testing into your spec sheet and inspection checklist.

An answer of "It's heat-resistant glass" or a quick check with a splash of boiling water is not enough. Set the glass type, coefficient of thermal expansion, required temperature difference and test method in numbers in the spec sheet, and test samples drawn from production lots. That is how you prevent breakage claims after sale.

1. Why Glass and Ceramics That Passed Inspection Break After Sale

Thermal shock resistance is a product's ability to withstand a sudden temperature change without breaking. When the surface of a cold container suddenly heats up, the outside tries to expand while the inside stays put, and that creates stress. If the stress exceeds what the material can take, it cracks or breaks. The outcome is already decided at the raw material, forming, annealing and firing stages. Breakage after sale usually comes down to these six causes.

• Glass type substitution: Borosilicate glass has a coefficient of thermal expansion of about 3.3×10⁻⁶/K, roughly a third of soda-lime glass (about 9×10⁻⁶/K). The same temperature change creates far less stress.
• Insufficient annealing: If the slow cooling of formed glass (annealing, 退火) is cut short, stress stays locked inside. Such products break from small temperature differences or light impacts.
• Uneven thickness: If the base or corners are much thicker, the thick and thin parts heat at different rates and crack along the boundary.
• Underfired ceramics: Cutting kiln temperature or time leaves the body (素地) underfired. It absorbs more water, and the expansion mismatch between glaze and body grows, causing crazing and breakage.
• Glaze-body mismatch: If the glaze's expansion doesn't match the body, the piece looks fine at first but hairline cracks appear on the glaze after a few dishwasher cycles.
• Overstated use claims: Labeling a microwave container "oven and stovetop safe" sends customers into conditions it would never pass in a test.

Thermal shock should be looked at separately from food contact safety. Food contact criteria such as lead and cadmium migration are covered in Episode 64, and drop testing to prevent breakage in transit in Episode 127. This article focuses on how well the product withstands temperature change.

2. Test Types — Temperature Difference Tests vs Use-Condition Tests

If you only ask for "a heat test," the factory may simply pour boiling water in or pick whatever conditions suit it. The same product can give very different results depending on heating temperature, cold water temperature and the number of cycles, so define the method and conditions first.

Read ΔT Together With the Conditions

A result of "passed ΔT 120°C" means different things depending on the hot and cold temperatures, how many cycles were run and how many samples were tested. Moving a piece from a 140°C oven into 20°C water and moving it from 100°C water into a -20°C frozen state are not the same test, even if the number matches. Don't write only the temperature difference in the spec sheet; include heating temperature, cooling temperature, hold time, number of cycles and sample size.

Coefficient of Thermal Expansion, Not the Word "High Borosilicate"

Chinese factory quotes often say "高硼硅" (high borosilicate), but the term is sometimes applied to low-boron medium borosilicate or even soda-lime glass. Borosilicate glass 3.3 has a coefficient of thermal expansion of about 3.3×10⁻⁶/K, and a lab can tell the difference fairly easily by measuring it. At the quotation stage, get the expansion coefficient and the raw glass supplier instead of a name.

3. What Criteria to Set

Here are approximate ranges commonly referenced in the industry. Requirements differ by buyer and market, so use the table only as a starting point and set actual criteria based on product shape, use claims and the lab's advice.

For ceramics, look at water absorption along with temperature difference. An underfired body absorbs more water, so it may pass the first test but then craze and eventually break after going back and forth between the dishwasher and microwave. Porcelain is commonly held to 0.5% or less water absorption; stoneware can be allowed a higher value, but write that number in the spec sheet.

4. Why You Shouldn't Take Factory Reports at Face Value

When asked for test reports, glass container and ceramics factories often send documents saying "all tests done," but these documents have a few gaps.

5. Common Supplier Tricks

Material differences in glass and ceramics barely show on the surface, which makes substitution easy. Here are the types we have actually encountered in consultations.

The most common case, as in the opening example, is making only the approval samples in good material. Borosilicate and soda-lime glass have similar weight and clarity, so an inspector can't tell by picking them up. Keeping a few approval samples aside and testing them alongside production samples under the same conditions catches many cases. Plastic resin substitution is covered in Episode 125.

6. What Happens When Products Break From Heat

Thermal shock failures don't show at shipment; they appear once customers start using the product, and the damage is larger than with other defects.

• Burns and cuts: When a container of hot food breaks, it can cause burns and cuts from glass shards, which becomes a product liability issue.
• Sales suspension and recalls: Once it becomes a safety issue, sales channels may delist the product and demand a recall.
• Labeling and advertising issues: If claims like "heat-resistant glass" or "oven safe" are untrue, they can be treated as material misrepresentation.
• Whole-inventory defects: Products from the same lot or kiln carry the same risk, so it is hard to sort out just the bad ones.

For product liability and recalls, see Episode 53; for after-sales returns management, see Episode 55.

7. What to Include in the Spec Sheet

A single line saying "heat-resistant glass" gets interpreted differently by every factory. Put the items below into a table and attach it to the spec sheet. For how to write a spec sheet, see Episode 59.

Asking for raw glass delivery records may seem excessive, but it does more than testing. Tests only check a few samples, while delivery records show whether the specified glass actually came in for the full production quantity. For change management, see Episode 119.

8. How to Add Thermal Shock Testing to Your QC Checklist

Sample Stage

When you receive approval samples, set a few aside and send some to a third-party lab for expansion coefficient and thermal shock testing, separate from the factory's report. These results become your baseline for comparing production.

Production Stage

During production, pull products from each lot as they leave the annealing lehr, check stress with a polariscope and run a water quench test. For ceramics, take samples from each kiln load and measure water absorption. Get test records for every lot, but if the results read the same every time and there are no photos, start by questioning whether the tests were actually run. For production scheduling, see Episode 102.

Pre-Shipment Inspection Stage

The inspector picks packed products personally, measures weight and thickness, and checks that base embossing and mold numbers match the approval sample. Run a simple quench test on site with boiling and cold water, and send leftover samples to the lab to confirm the expansion coefficient. For pre-shipment inspection, see Episode 100.

9. Sample Request Wording for the Factory

Putting the wording below in your purchase order or spec sheet takes material choice out of the factory's discretion and makes it easier to assign responsibility if breakage claims arise.

"本产品须使用与确认样相同的高硼硅3.3玻璃，线膨胀系数为3.3±0.1×10⁻⁶/K，原料玻璃供应商须与确认样一致，克重允许偏差±5%。耐热冲击按ISO 7459测试，温差120℃时全部样品不得破裂；-18℃冷冻状态放入预热至220℃的烤箱，不得破裂。未经我方书面同意不得更换玻璃材质、原料供应商或生产工厂，如需更换须事先书面通知并重新测试。每批须进行应力检查和冷热冲击测试，并提交注明品号、批号、测试方法、结果及照片的测试记录和原料玻璃出货凭证。测试样品由我方验货员随机抽取。如有不合格，须加抽复测；复测仍不合格的批次暂停出货。"
(This product must use the same borosilicate 3.3 glass as the approval sample, with an expansion coefficient of 3.3±0.1×10⁻⁶/K; the raw glass supplier must be the same as for the approval sample, and the weight tolerance is ±5%. Thermal shock is tested per ISO 7459, and no sample may break at a 120°C temperature difference; the product must not break when moved from a -18°C frozen state into an oven preheated to 220°C. The glass material, raw material supplier or production factory may not be changed without our written consent; any change requires prior written notice and retesting. Each lot must undergo stress inspection and quench testing, with test records showing item number, lot number, test method, results and photos, plus raw glass delivery records. Test samples are drawn at random by our inspector. If a sample fails, additional samples are drawn and retested; any lot that fails again is held from shipment.)

10. Three Real Cases

These are anonymized versions of types we see often in consultations. Amounts and figures are examples.

Case 1 — Borosilicate Glass Storage Containers

This is the opening case. The brand took the test results to the factory, but the spec sheet only listed "heat-resistant glass," capacity and weight, with no glass type, expansion coefficient or temperature difference criterion. The factory held its ground, saying "it is heat-resistant glass." In the end the factory admitted using glass different from the approval sample and refunded part of the amount, but recalls, refunds and product page revisions still left a loss of about 38 million won. From the next order on, the brand puts the expansion coefficient and raw glass supplier in the spec sheet and sends samples from every lot to a lab.

Case 2 — Ceramic Mugs

A café merchandise company buying mugs from a ceramics factory in Chaozhou, Guangdong, received reviews three months into sales saying "after a few runs through the dishwasher, hairline cracks appeared, and when I poured hot coffee in, it split below the handle." Returned mugs had water absorption above 3%, against 0.3% for the approval sample. The factory had shortened the firing time to meet volume. The company now writes a water absorption criterion into the spec sheet and checks samples from each kiln load.

Case 3 — Double-Wall Glass Cups

A company importing double-wall glass cups from a glass factory in Zhejiang received several complaints that "the inner cup cracked when I poured hot water in." The material was indeed borosilicate, but a polariscope showed heavy stress left at the rim where the inner and outer cups join. The factory had cut annealing time during peak season. The company now requires stress check photos for every lot. If you check only the material and not the annealing, the same thing happens again.

11. Common Mistakes

• Trusting words like "heat-resistant glass" or "high borosilicate" without setting expansion coefficient and temperature difference in numbers
• Judging by a single splash of boiling water
• Accepting a raw glass supplier's report as a finished-product report
• Leaving the uses on the product page (freezer to oven, dishwasher) out of the test conditions
• Not checking water absorption and crazing on ceramics
• Checking only the material and ignoring annealing and residual stress

12. Thermal Shock Control Checklist

Specification Stage

• Glass type, expansion coefficient and raw glass supplier are set
• Temperature difference, heating and cooling temperatures, cycles and sample size are written down
• Every use condition in the sales copy is included as a test item
• For ceramics, water absorption and crazing criteria are written down
• Per-lot sampling and handling of failures are defined

Sample and Production Stage

• A few approval samples are kept aside
• Approval samples were tested separately at a third-party lab
• Per-lot stress check and quench test records were received with photos
• Raw glass delivery records were checked

Pre-Shipment Inspection Stage

• Weight and thickness of production units were measured
• Base embossing and mold numbers match the approval sample
• An on-site quench test was run and leftover samples were sent to the lab

Wrapping Up - Check Temperature Difference and Material, Not One Splash of Hot Water

At room temperature every glass container looks equally solid. The difference shows only when a customer takes one out of the fridge and puts it in the oven. Writing the glass type, expansion coefficient and required temperature difference into the spec sheet and checking test results for every lot doesn't cost much. Put a few hundred thousand won in testing next to tens of millions in recall costs, and it is clear which is cheaper.

There is something you can do right now. Ask the factory currently producing for you to send the expansion coefficient of the glass in use, the raw material supplier and thermal shock test records for recent lots. If the answer is a single line saying "It's heat-resistant glass," draw up a thermal shock control sheet and send it before your next order.

StartMade supports you on the ground with spec sheets covering material and thermal shock criteria, third-party testing of approval samples, checks on production lot stress inspections, and pre-shipment inspection. If you've already received breakage claims, we start with testing remaining stock and finding the cause; if you're preparing a new product, we start with material selection and setting criteria.