Gadget Guts: Testing lithium-ion batteries

With the recent news that lithium-ion batteries in smartphones are exploding and catching fire, we decided to test a variety of lithium-ion batteries to determine whether they do indeed catch fire and/or smoke when punctured.

To this end, we tested four different battery types. In addition to using a drill press (as the video below shows), we tested the batteries (not shown) using a pocket knife and an ax. We found that the drill press provided the best immediate results, the safest environment, and optimal conditions for video recording.

Battery Types:

  1. We intentionally withheld manufacturers’ information.
  2. We charged all batteries prior to testing.

As you may have noticed, we took steps to protect ourselves using personal protective equipment (PPE), including leather gloves, a Z87+ full face shield, and/or Z87+ safety protection eye glasses.

Battery #1:

  • Physical size: 18650
  • Voltage rating: 3.7 V
  • Energy rating: 4,000 mAh (milliampere-hours)

Battery #2:

  • Physical size: custom for specific camera brands
  • Voltage rating: 7.2 V
  • Energy rating: 2,200 mAh

Battery #3:

  • Physical size: 18650
  • Voltage rating: 3.7 V
  • Energy rating: 2,500 mAh

Battery #4:

  • Physical size: 21 x 34 x 113 mm
  • Voltage rating: 11.1 V
  • Energy rating: 2,100 mAh 

Battery Testing Conclusions:

  • One battery brand/type (battery #1) yielded no fire and little to no smoke.
  • Two batteries (#2 and #4) either smoked moderately to heavily and/or displayed the boiling of internal chemicals.
  • Only one battery type (#3) resulted in violent flames and fire.
  • We don’t know why some batteries catch fire while others merely smoke and yet others still seem ultra-safe when punctured. But we venture to provide some educated-guesses:
  • some batteries have built-in safeguard protections
  • some batteries do not have the energy ratings that their advertised ratings suggest
  • some battery manufacturers use more lithium, or less water, than other manufacturers use
  • Some notes about lithium (Li):
  • Lithium is a soft, silvery metal that reacts vigorously with water
  • Lithium carries atomic #3 on the periodic table
  • Lithium is an Alkali metal
  • Alkali metals become more vigorous with water as you move down the Alkali metal group on the periodic table




Read more at the  Special Report: Beyond the exploding battery



Power Supplies & Energy Storage Technology

Related Posts


Join the conversation!

Error! Please fill all fields.

  • Maximilian Teodorescu

    Fun project. Haven’t really seen anyone else do this kind of thing.

  • John Warner

    Very interesting video, I would like to add a couple of comments:

    First a couple of comments on safety:
    1) Using a drill press to penetrate a lithium-ion cell is extremely dangerous. Cylindrical cells have a vent in the cap of the cell, so there is a huge risk that they could vent towards the engineer.
    2) One of the “standard” tests that is performed at various levels (cell, module, pack) is a nail penetration test. In this test a straight “nail” is used and is inserted at a constant rate into the cell (see USABC testing guide, SAE J2464 or Chinese GB/T 18384). The big difference here is the turning motion of the drill which will “tear up” the electrodes internal to the cell and will very likely cause an internal short circuit of the cell vs a straight insertion which will also cause an internal short but in a “safer” manner.
    3) Time – it can take a lithium-ion cell up to an hour or more to go into “thermal runaway” (fire, smoke, etc, so those cells should have been placed into a safety enclosure, explosion proof box after the test. Remember the Chevy Volt Rollover test that NHTSA did? It took the battery a week in the vehicle to go into thermal runaway (due to a leak in the liquid cooling system which caused a short).
    4) Location – lithium ion cell testing should be conducted in a facility designed for it. Any abuse testing is typically done in an explosion proof room or chamber to ensure the protection of the testing staff and engineers.
    5) The gases that escape from a lithium-ion cell during a venting event are hazardous – so doing it outdoors certainly helps but should be done somewhere where the tester isn’t exposed to these gases.

    Ok, with the safety disclaimers out of the way
    • Penetrating a lithium-ion cell, regardless of whether it is a cylindrical, pouch or can cell will very frequently result in a thermal failure event. In effect what you are doing here is introducing an internal short circuit into the cell and just like any other circuit this will result in a failure.
    • Different chemistries will have different “failure” states – LCO, NMC and NCA chemistries are higher energy and will be more likely to fail in this manner. While LFP and LTO are lower energy density cells and are less likely to fail in as violent a manner. But virtually all chemistries can be forced to fail.
    • Voltage – one of the things that is very important in this type of test is what voltage level the cells are at. If the cells were fully charged they will be more likely to go into a thermal event (more energy to be release at high states of charge). If the cells are at a low state of charge they will be less likely to fail in this manner (less energy to be released). So depending on what the actual state of charge those cells were at the test could be misleading. For example, were the two cells that did not flame up at a low state of charge (fully discharged)? and what chemistries were they?
    • One other comment about voltage is that it appears that the test included two cells and two “packs”. The nominal voltage of a Li-ion cell is from 3.3V (LFP) to about 3.7V (NMC, NCA). so a 7.2V pack would consist of two (2) 3.6V cells in parallel; and an 11.1V pack is really three (3) 3.7V cells in parallel. So the results aren’t entirely an apples to apples comparison as there is a lot more energy to be released in the packs than in a single cell.

    • steve sparks

      John your last paragraph should say series not parallel

    • Nick Davis

      Hi John,
      Thank you for your feedback and questions.

      The spec sheet we received with battery #1 did not reveal the chemistry type. However, based on the nominal voltage of 3.7V I’m assuming it’s an NMC type battery.

      Similarly, the chemistry type of battery #2 was not given. But again, based on the nominal voltage of 3.6V (7.2V / 2) and because this battery is designed specifically for cameras I’m guessing it’s an LCO type.

      Also, all batteries were fully charged prior to testing.

    • Paul O'shea

      Thank you very much for your comments and explanations. These are excellent observations that will help all of us learn something. I especially like the safety comments and #3 about thermal runaway.