Is Donut Labs Battery Lithium Ion?

---

Support CleanTechnica’s work through a Substack subscription or on Stripe.

---

A claim was made in a youtube video by the internet handle Ziroth.

Before we get deep into this, I need to explain lithium-ion battery charging. Lithium batteries are charged with a voltage source that is current limited. Confusing? Bear with me. When the battery starts at lowest voltage, the difference between source voltage and battery voltage is large and the source is current limited. The current is the same while the output voltage is changing. Therefore, it is being charged by a current source. Once the battery gets close to the final voltage, the source current starts falling until it gets very low. Then it behaves as a voltage source. You can see it in the voltage and current graphs from VTT in Figure 2. Click on the links here to get that.

Here are the curves shown by the Ziroth video above:

He compared a curve that shows voltage versus SOC with a curve that shows voltage versus time. It is bad practice to compare two curves with different parameters. Two curves should be viewed with the same units on both scales. Why? It is general scientific practice. Not yet a crucial problem, but it can sometimes lead to false conclusions. The second curve appears to be from the downloadable VTT report, Figure 2. Its horizontal units are labeled hr, and the text matches hours. Limits are still somewhat observable, more on that later. The second curve shape changes quickly at first, rising suddenly, then slows. What do we know about batteries at low SOC? They charge fast at first, then they slow down. The lithium curve shows the range of cell voltages versus SOC.

The first curve is reputed to be lithium-ion by Ziroth, and does look like typical NMC. Now look at the next curve. It is the Donut Labs curve. A lithium battery voltage cannot go too low or too high. If it goes too low for too long, it gets bricked. The battery cannot be used again. If it goes too high, an NMC battery will be damaged or catch fire. Both min and max must be limited. That is, lithium must be limited on the lower limit. Some sources say the particular sodium-ion battery can be fully discharged to zero. The charger does the limits, not the battery. What we get from the curve independent of the charger is the steep rise at the beginning of the charge. That rise tells us we are near 0% SOC. The upper voltage on the right curve is as flat as a board. That upper voltage was limited by the charger, not the battery. When that looks flat, the charging is not increasing battery voltage much any more, but we do not know where we are in an SOC graph. The VTT spec says stop charging when nearing that upper voltage when the current is a given amount.

“The cell was charged with a 24 A current until the highest recommended voltage of 4.15 V was reached, followed by a constant-voltage charge with 4.15 V until the current decreased to 1.2 A.”

However, we can get an idea about what the battery characteristics are like from those limits. We can see the approximate full discharge 0% SOC voltage. It is somewhere in that inflection point from vertical to where it starts to go more horizontal, but that point is not accurately determined. We also cannot tell the exact point where 100% SOC happens. Worse yet, the time to 100% SOC has a lot of variability, because the smaller we make the current limit, the more time it takes. Remember, this is where the current falls slowly. The closer to 100% SOC it gets, the longer time it takes. It’s an exponential relaxation curve like the discharge of a capacitor with resistor. It has a long tail. Because of this, 100% SOC is not a good metric to determine time of charging. The time it takes to stop charging when current decreases to 1.2A could be more if it stopped at 0.12A and it is not linear. That is why I think the statement 5 minutes to 100% SOC is not a real engineering specification. It lacks a condition. 

That leads us to one more thing. There were many, many statements by more than one person in the Donut Labs video interviews I saw. Which ones are official specs? I don’t know what rigor is made in those statements compared to an engineering specification. That sounds trivial, but it is not. Conditions matter when testing and specifying a product. Without getting tangled up in it too much, I say only the time from 0 to 80% SOC makes any sense to me for 11C testing. That is because the 100% SOC end point is inaccurate and is not specified what the end point is. The 0% SOC point is easier because it does not depend much on time or voltage, only that the charge is near zero. It only requires that the battery be drained. It makes little difference to the charging time measurement exactly how well it is drained. The time to 80% SOC point is relatively well known and we can estimate accuracy. Really, what we want to know is how fast it charges to a reasonably accurate amount of charge, not a perfectly accurate amount. If the charge estimation was off by 5%, the time is only off about that much while the current is constant. That is also because most or all of the charge curve is linear up to that point. It is only after the charge curve is mostly non-linear that the charge time becomes less accurate or predictable. 

Therefore, 100% SOC charge times are not a reasonable method of ascertaining max charging speed or charge rate, C. The spec should be made to a level of SOC other than 100%. If Donut Labs said that, it was not a proper engineering spec. No matter, because the real performance can still be ascertained.

Below is a screenshot of a Ziroth video about CATL’s reinforced sodium-ion battery, a completely different battery chemistry. I wrote about that battery in a recent story.

Look familiar? This one shows SOC reversed, starting from full voltage and discharging toward 0% SOC. The limits are similar if not the same, but it is a sodium-ion battery. They call it a reinforced sodium-ion battery, and it is from a patent I wrote about coming from CATL. They deliberately doped the cathode of a sodium-ion battery to change the curve. It worked. Do you see a major difference in voltage limits? I don’t. 

He does say, “this curve is like an electrochemical fingerprint for the cell, and it isn’t something you can just change around depending on the application, at least not from anything I’ve ever seen.” 

Uh-oh. That last statement is an exception. I do agree with him. I never saw two chemistries with nearly identical charge curves before. Until now. 

That was at 3:02.

Let’s go back to the Ziroth video. There are too many statements to cover them all. The statement is made, “I now believe it is extremely unlikely to be true and I give it a naught point one chance of being true.”

“I can say with extremely high certainty, that what we are looking at is a lithium-ion battery.” (at about 2:00)

From that, a whole series of conclusions are made by Ziroth.

As I said, I am loathe to draw a string of conclusions about other matters unrelated to charge time from this limited information. I would make absolutely certain the first statement is true before I dared do that.

Then there is a statement, “Charging times to full are also significantly slower than previously stated.”

As I said before, I do not consider statements made by Donut Labs during CES to be be engineering specifications. I do not have an engineering specification. Before a target is tested to see if it is met, I have to know precisely what that target is. 100% SOC is not a precise target without an ending current cutoff limit. It is common for semiconductor specifications to include statements of performance that do not have conditions that may be achieved, then in the body of the text, show conditions and guaranteed results. Those are not the same. The independent tester really cannot do accurate tests to 100% SOC. So they showed times to 80% SOC and C rates over that range. The charge to 80% is 4.5 minutes. The charge rate C, in 4.5 minutes to 80%, gives the C rate. If the charge was 100% in 5 minutes, the C rate would be 60/5, or 12. If the charge is less by 0.8 and done in 4.5 minutes, we get 60/(4.5/0.8) = 10.667. The charge rate increases with temperature. The temperature was not specified. It looks as if it might reach 11C if the temperature rises a little higher. 

Based on my experience, manufacturers do not always give accuracy limits for C rate measurements. This meets typical standards. Sorry, the spec we started out with is not a rigorous engineering spec with accuracy limits stated. The test is adequate to support the claims of high charging speed within reasonable accuracy for the time being. 

There are many statements that go beyond charging speed. Since the tests only are designed to measure charging, I refrain from making any statements about other performance metrics based on the tests. Without proper measurements, specs, and conditions, I cannot make firm conclusions. There are more statements, like ones about red flags I will not include. Red flags are not proof. They may mean something, but must be weighed much more carefully and on their own, and cannot be counted on as proof. 

There is a series of statements about cell capacity attempting to determine that. He uses a display battery at CES. Then he says, “Sure they could be testing different cells.” There is no sense in going further. There is no way to  know that the cell presented to independent tests is the same one used in the display. It’s speculation. Then he says he thinks it being a different cell is unlikely? That is not accurate. There should be no conclusion from lack of knowledge.

The cell was stopped from exceeding 90C. The limit was assumed set by the battery. When you say, “I don’t know why there would be,” it means you do not know what the reason was it was stopped. There is no conclusion from lack of knowledge. Just because the battery test stopped at 90C does not mean the battery has a safety limit at 90C. It might, but the conclusion cannot be made from that yet.

It goes on.

Cells seen expanding? I see something change in some foil across the two tabs. The rest of the battery underneath the foil, I see no movement. There were no measurements of expansion that could be referenced. The statement “I see expansion” is not scientific or accurate. The cell cannot be seen below the foil and there are no efforts to gauge the cell expansion properly. Does not mean it expanded or not. From this test, there is no way to determine how much expansion there is. 

I am going to stop there. I think Ziroth seems to be a reasonably nice, intelligent person. I have no intention of being personal here. It’s not about persons. It is about ideas, and this notion Donut Labs proffers about skepticism. I have my own ideas about engineering and rigor and I have my own conflicts about that, and hold to high standards. When I appear to criticize, keep in mind that I come by it honestly, and it is not about people, but about my excessive desire for accuracy and certainty. Being a design engineer does not always make me the life of every party, at least while I am in that mode. When others make statements or do things that are inaccurate or unsubstantiated, I get antsy. I cannot abide by that kind of uncertainty, so I go to ridiculous extremes to find out. That bugs people, and I know it. Please accept my apologies in advance if my excessive engineering bothers you.

It’s just that others have picked up on this video and then said, “see, Donut Labs’ claims are false.” I cannot absolutely come to conclusions based on limited data that does not absolutely confirm or deny a conclusion. If it does, I will state it firmly. As you can see, I do not think those blurbs and images showing things like minutes to 100% SOC are true engineering specs. If they were, you would see conditions and limits fully specified. A test engineer must have that kind of spec to make a measurement. That is why the independent tester stated 0 to 80% SOC in 4.5 minutes. 80% SOC is determined from the amount of charge from the point it was charged to max voltage to a discharge point where the current dropped off to some limit. He made it an engineering spec. He had to. His test machine demanded the settings because they are real-world settings. Likewise, C rate is not always a tightly specified value with percentage variation limits. I can only say the cell presented to independent testing delivered the performance VTT stated. VTT made their statements. In the video, they also added that the Donut Labs device was better than other solid-state devices they tested because it did not require special clamping methods. 

Understand that variation and limits are not given before production in anything but general terms, and if there is something released, it is always accompanied by something that says it is pre-production or the like. That is because those specifications determine yield. You cannot do yield until after the product goes into factory production. You can try for tolerances, but cannot predict them exactly. It depends on both process and design.

I can stop now. It is time for a pause and reflection. Truth is not that simple. I cannot make a statement saying Donut Labs failed charging speed tests. There is enough evidence that it performs like its statements say. I would prefer a true engineering spec, but at this level of development, an engineering specification does not make sense. The tests are adequate to determine the truth of the statements for now. It is enough to say Donut Labs’ statements are meaningfully and reasonably correct and accurate enough for now. I can also say that statements concluding that this is a lithium-ion battery are questionable. I will not go further than that except this: If there are two curves that match and the chemistries are different, then Donut Labs’ battery might not be a lithium battery. At least I can say it might be the other curve that seems to match VTT’s test results. Those curves do match a reinforced sodium-ion battery as detailed in a CATL sodium-ion patent and the curves were proffered by Ziroth. But Donut Labs battery energy density is more than the sodium ion. Isn’t that curious. I won’t put $5,000 on it. If you make me do that, I may study it and never come back again. I don’t know anything for sure yet, but it sure is curious. Interested yet?