Glycolic acid carbon coating for lithium-rich manganese cathodes

A mild glycolic acid wash plus a thin carbon skin lifts the first-cycle efficiency of a lithium-rich cathode from 83.5 to 89.9 %. We check the capacities, fade and impedance with fibub's battery and EIS tools.

Lithium-rich manganese oxides are among the highest-capacity cathode materials for lithium-ion batteries, storing well over 250 mAh of charge per gram. Two problems hold them back: a lot of lithium is lost in the very first charge, and the voltage and capacity sag as the cell is cycled. A 2026 study in Batteries treats the cathode powder with a mild glycolic acid wash and a short bake, which rebuilds its surface and leaves a carbon skin a few nanometres thick. This article shows how the battery cycling and the EIS fitter let anyone check the paper’s efficiencies, fade and resistances, and where the numbers tell a more mixed story than the headline.

The paper in brief

The team made a lithium-rich cathode (Li-, Mn-, Ni- and Co-based layered oxide) and stirred it in ethanol with 3, 5 or 7 % glycolic acid, then baked it at 450 °C. The acid takes some lithium and oxygen from the surface, which turns the outermost layer into a spinel-like structure that lets lithium in and out more easily; the leftover organic acid turns into a thin carbon coat. The 5 % sample (called LRM-G5) worked best.

They report a higher first-cycle efficiency (89.9 % against 83.5 %), more capacity at every rate in their rate test, slower voltage decay (1.7 against 3.5 mV per cycle), a lower charge-transfer resistance, and a lithium diffusion coefficient about five times higher, measured by impedance.

Checked with the tools

Step 1: the first cycle. A cathode’s first-cycle efficiency is the discharge capacity divided by the charge capacity. Working back from the paper’s efficiency and first discharge, the Battery cycling analyser gives a first charge of 310.4 mAh/g for the untreated cathode and 317.9 mAh/g for the treated one. So the untreated cathode loses 51.2 mAh/g for good in the first cycle and the treated one only 32.1 mAh/g: about 37 % less lithium wasted. That is the clearest gain in the paper.

Step 2: rate capability. Dividing each discharge capacity in the paper’s rate test by the 0.1 C value gives the share kept at each rate (Figure 1). At 3 C (a full discharge in 20 minutes) the treated cathode keeps 64.5 % against 60.5 %, and gives 183.9 against 160.2 mAh/g. At 1 C, though, the shares are close (77.4 against 79.1 %): the treated cathode gives more capacity because it starts higher, not because it holds up better at that rate.

Rate capability: share of the 0.1 C capacity keptRate capability: share of the 0.1 C capacity kept607080901000123C-rateCapacity kept (%)Glycolic-acid treated (G5)Untreatedfibub.com
Figure 1. Share of the 0.1 C capacity each cathode keeps at higher rates, worked out from the discharge capacities in the paper's rate test.

Step 3: fade at 1 C. The paper’s main text gives the untreated cathode 81 % of its capacity after 100 cycles at 1 C and the treated one 76 %, from first capacities of 253.9 and 239.2 mAh/g. We simulated 100 cycles with a straight-line fade between those points and ran them through the analyser: 0.192 % per cycle untreated, 0.242 % per cycle treated, and 80 % reached at about cycle 84 for the treated cathode (about cycle 105, extrapolated, for the untreated one). The paper’s abstract says 81 % for the treated cathode, and the text says it keeps about 10 percentage points more after 200 cycles, so the 100-cycle numbers in the text and the abstract do not agree. Our check cannot settle which is right; it shows why readers should look at the cycling plot itself and not one number.

Step 4: impedance. The paper fits its impedance spectra with a solution resistance Re and a charge-transfer resistance Rct. We built spectra with the Impedance fitter’s Randles circuit from those values (Figure 2) and fitted them back: the fitter recovers every value exactly, as it should for clean data. With the 1.44 cm² electrode, the fitter’s exchange current density (how fast lithium crosses the surface at rest, i₀ = RT/(nF·Rct·A)) is 1.03 × 10⁻⁴ A/cm² for the fresh untreated cathode and 1.19 × 10⁻⁴ A/cm² for the treated one. After cycling, the untreated cathode’s Rct rises by 94 % (i₀ halves to 5.3 × 10⁻⁵ A/cm²), while the treated one’s rises by under 1 %. That is the strongest sign that the new surface protects the cathode.

Impedance after cycling (simulated from the paper's fitted values)Impedance after cycling (simulated from the paper's fitted values)01002000100200300400500Z′ (Ω)−Z″ (Ω)Treated, cycledUntreated, cycledUntreated, freshfibub.com
Figure 2. Nyquist plots simulated with the Impedance fitter's Randles circuit, using the paper's fitted Re and Rct and our assumed double-layer and Warburg values. An illustration, not the measured spectra.

Step 5: diffusion. The paper’s two diffusion coefficients differ by a factor of 5.3 (2.47 × 10⁻¹⁶ ÷ 4.65 × 10⁻¹⁷), which the paper calls about an order of magnitude. Because the Warburg coefficient σ goes as 1/√D, the treated cathode’s low-frequency tail in our simulation is 2.3 times less steep for the same assumed electrode.

Check Paper Tool result
First charge, untreated / treated (from efficiency and discharge) 310.4 / 317.9 mAh/g
Lithium lost in cycle 1 (not stated) 51.2 / 32.1 mAh/g
Share kept at 3 C (from the rate test) 60.5 / 64.5 %
Fade at 1 C, 100 cycles (main text) 81 % / 76 % kept 0.192 / 0.242 % per cycle
Rct rise after cycling 174.0 → 338.4 Ω; 149.5 → 150.7 Ω +94 % / +0.8 %
i₀, fresh electrodes (1.44 cm²) (not stated) 1.03 / 1.19 × 10⁻⁴ A/cm²
Diffusion coefficient ratio “about one order of magnitude” 5.3 times

How we made the graphs

Everything above comes from scripts/blog-figures/glycolic-acid-carbon-coating-lithium-rich-cathode.mjs in fibub’s public code. It calls:

  • cycling and cyclesTo from src/lib/sci/battery.ts, the maths behind the battery cycling, for the first-cycle loss, the fade per cycle and the cycle at which 80 % is reached.
  • MODELS (the Randles circuit), fitEIS and exchangeCurrent from src/lib/sci/eis.ts, the maths behind the EIS fitter.

Where the inputs come from: the efficiencies, capacities, Re, Rct, diffusion coefficients, electrode area and frequency range are from the paper. The straight-line fade, a coulombic efficiency of 99.6 % in the simulated cycles, the double-layer values (Q = 2 × 10⁻⁵, α = 0.85) and a Warburg coefficient of 300 Ω·s^−½ for the untreated cathode are our assumptions. Figure 1 is worked out from the paper’s capacities; Figure 2 and the 100-cycle fade are simulations, not the measured data.

How we test the maths

On every change to fibub, tests/sci.mjs checks these functions against known answers:

  • A straight-line set of cycles losing 1 % per cycle must give a fade of exactly 1 % per cycle and reach 80 % at cycle 21; the built-in example must give a first-cycle efficiency of 88 % and reach 80 % at cycle 184.
  • Every circuit in the impedance fitter, including the Randles circuit with Rs 20 Ω, Rct 100 Ω, α 0.9 and σ 30 Ω·s^−½, must be fitted back to its exact values, and a two-arc circuit must be flagged as unreliable on one-arc data.
  • The exchange current density for 1 Ω·cm² at 25 °C must be 0.025693 A/cm² (RT/F).

Check it yourself

The first cycle, on the battery cycling:

  1. In Cycle number, charge capacity and discharge capacity, one cycle per line, type two lines: 1 317.9 285.8 and 100 182.5 181.8 (the treated cathode’s first cycle at 0.1 C, and our simulated cycle 100 at 1 C; the retention it shows mixes two rates, so read only the first-cycle line here).
  2. Set Capacity unit to mAh/g, Coulombic efficiency is to discharge ÷ charge (cathode, full cell) and Reference cycle for retention to 1.
  3. First cycle reads a first-cycle efficiency of 89.9 % with 32.1 mAh/g lost. Change the first line to 1 310.4 259.2 for the untreated cathode: 83.5 % and 51.2 mAh/g.

The impedance, on the EIS fitter:

  1. Paste this simulated spectrum of the cycled, untreated cathode (frequency in Hz, Z′ and Z″ in Ω) into Frequency and impedance, three columns, one frequency per line:
100000	12.1	-0.5728
31600	12.33	-1.522
10000	12.98	-4.034
3160	14.87	-10.62
1000	20.96	-27.4
316	43.03	-65.81
100	115.6	-124
31.6	242.7	-138.2
10	335.3	-106.8
3.16	392.6	-99.84
1	455.2	-134.5
0.316	552.9	-219.6
0.1	720.6	-381.5
0.0316	1016	-674.2
0.01	1540	-1197
  1. Leave Columns are on f, Z′, Z″ and Equivalent circuit on the R(Q(RW)) Randles circuit.
  2. Type 1.44 in Electrode area (cm², optional), 1 in Electrons n and 25 in Temperature (°C).
  3. The fit reads Rs about 11.96 Ω, Rct about 338.4 Ω, Rct × area about 487 Ω·cm² and Exchange current density i₀ about 52.7 µA/cm².

What researchers can use it for

  • Turn efficiencies into lost capacity: the battery cycling shows how many mAh/g a coating saves in the first cycle, which matters more for a full cell than the percentage.
  • Compare fade fairly: paste cycling data from two cells and read fade per cycle and cycles to 80 % on the same terms.
  • Re-fit impedance with several circuits in the EIS fitter and see which values the data can really pin down.
  • Turn Rct into i₀ so cells with different electrode areas can be compared.
  • Teach why the first cycle of a lithium-rich cathode loses so much lithium, and what a growing impedance arc means.

The fine print

  • The fade in Step 3 is a straight line between two stated points; real cells fade unevenly, and the paper’s 200- and 400-cycle tests are not modelled here.
  • The text and abstract disagree on the 100-cycle retention; we used the text’s values and say so.
  • i₀ from Rct assumes one simple charge-transfer step near equilibrium; a real cathode surface has several processes in the arc.
  • The double-layer and Warburg values in Figure 2 are assumed, so only the arc widths (Re and Rct) follow the paper.

Sources

  • Yang, X., Miao, J., Chen, Y., Fang, Y., Wang, H., Peng, G. (2026), Glycolic Acid-Induced Surface Reconstruction and In Situ Carbon Coating for High-Electrochemical-Performance Lithium-Rich Manganese-Based Cathodes, Batteries 12(2), 70, doi:10.3390/batteries12020070. Licence CC BY, checked on the journal page, 2026-11-16. Values used here: first-cycle efficiency 83.5 % (untreated) and 89.9 % (treated) with first discharge 259.2 and 285.8 mAh/g at 0.1 C; discharge capacities at 0.1, 0.2, 0.5, 1, 2 and 3 C; 1 C cycling from 253.9 and 239.2 mAh/g with 81 % and 76 % kept after 100 cycles (main text); Re and Rct before (7.167 and 174.0 Ω; 3.673 and 149.5 Ω) and after cycling (11.96 and 338.4 Ω; 8.476 and 150.7 Ω); lithium diffusion coefficients 4.65 × 10⁻¹⁷ and 2.47 × 10⁻¹⁶ cm²/s; electrode area 1.44 cm²; impedance from 100 kHz to 10 mHz at 25 °C.
  • Exchange current density from charge-transfer resistance, i₀ = RT/(nF·Rct): Bard, A. J. and Faulkner, L. R., Electrochemical Methods, 2nd ed. (2001), chapter 3.

If you are an author or publisher and want a credit or a passage changed or removed, tell us and we will act quickly.

Tools in this article

Read next