Peak Bifurcation of NMC: H1, H2, and H3 “Phases”
In March 2021, a Nature Materials paper came out that altered my brain chemistry. This post is dedicated to that paper and explains the controversy behind layered metal oxide phase changes with a focus specifically on NMC. Let’s get into it.
A very brief background and how we talk about structures
NMC (LiNiₓMnᵧCo₁₋ₓ₋ᵧO₂), LCO (LiCoO₂), and LNO (LiNiO₂) all share the same crystal structure, which belongs to the hexagonal R3̅m space group.
This structure is classified as O3 in Delmas’ notation, indicating a face-centered cubic (fcc) arrangement.
The letter "O" refers to the octahedral coordination of lithium (Li) within the structure.
The number "3" signifies the three transition metal oxide layers present in each unit cell.
The layers are arranged within the ab plane, while the c-direction is perpendicular to these layers, meaning that the unit cell parameters satisfy a = b ≠ c.
Phase Transformations and Notation
Most phase transformations in these materials occur without disrupting the O3 stacking sequence, meaning that the O3 structure remains intact. Because of this, a secondary notation system is used to describe different phases:
Phases are labeled using a letter and a number:
The letter represents the symmetry of the unit cell.
The number indicates the order in which that phase appears during charge.
Structures that belong to the rhombohedral lattice system are typically labeled as "H" for hexagonal.
Other notable structures include:
Distorted O3 structures, which are classified as monoclinic (M).
O1 structure, belonging to space group P3m1.
Hybrid H1-3 phases, which feature alternating O1 and O3 blocks.
Figure 1
Illustrations of the crystal structures relevant to the layered cathodes. Unit cells are shown in black (hexagonal) or blue (monoclinic) and Li, TM, and O atoms in light green, dark green, and red respectively. [1]
LCO
In 1998, LCO was identified (through simulations) to form three different phases during cycling:
the rhombohedral form of LCO (O3)
the hexagonal form of CoO2 (O1)
a hybrid of the previous two structures (H1-3) [2, 3]
This was confirmed experimentally in 2002 with XRD data.[4]
Figure 2
Schematic illustration of the three host structures O3, O1, and H1-3. Upper case letters describe the stacking of the close-packed oxygen layers.[2]
LNO
LNO was also found to have a phase change during cycling.[5-8] During charging, the c- lattice parameter increases while the a-lattice parameter decreases. This behavior is well established for LNO and isostructural NMC- and NCA-based cathodes: without the positively charged lithium ions in between the negatively charged metal oxide slabs, the like charges repel each other. Opposites attract and all that.[8]
At least four different Li1-xNiO2 phases have been described in literature: hexagonal H1, monoclinic M, hexagonal H2, and hexagonal H3 phases.[8] The sequence of phases observed for LiNiO2 is H1 → M → H2 → H3 with all phases having an O3 stacking sequence.[1]
Figure 3
Evolution of a- and c-lattice parameters from Rietveld refinement analysis of operando XRD data obtained for a Li/LNO cell [8]
Of all the known phase transformations occurring in LiNiO2, the H2 to the H3 phase at high state of charge (SOC: 4.15–4.25 V) is believed to have the most detrimental impact on the material’s intrinsic stability, as it is accompanied by a sudden collapse of the structure along the crystallographic c-axis.[8]
C-rates of LNO studies referenced
Lithium/vacancy ordering[5] = theoretical
New Findings[6] = C/13 and C/84 rates
Structural Characterisation[7] = C/200 rate
Phase Transformation behavior[8] = C/10 rate
NMC
NMC is a solid solution of LiCoO2, LiMnO2, and LiNiO2 of hexagonal R3̅m space group.[9] Therefore, it can be reasonably assumed that phase changes occurring in LCO or LNO during cycling should also appear in NMC. RIGHT???
In 2006, a synchrotron-based in situ XRD study was done on NMC111 and NCA, and they were found to follow similar structural changes as LNO during cycling: expansion of the c-axis and contractions along the a- and b-axis during charge, and a major contraction along the c-axis with slight expansions along the a- and b-axis near the end of charge at high voltage limit. The rate cycled was C/5 for those of us keeping track.[10]
Figure 4
Variations of the lattice parameters of NMC111 as a function of x during first charge. The hexagonal phase (H1) lattice parameters (a and c) are indicated with circles and squares, and the other hexagonal phase (H2) with upper and lower triangles.[10]
Despite significant evidence of the phase transformation occurring in LNO and LCO (and some evidence in NCA and NMC), there is a lot of disagreement among NMC studies.
Figure 5
In situ XRD of NMC111 (Li1−xNi1/3Co1/3Mn1/3O2) during the first charge. Contour plot of the 003 diffraction peak of NMC111 with increasing x between x = 0 and 0.7 during the first charge process at different C rates. [11]
Zhou et al: XRD patterns of the NMC111 cathode were captured during the first charge - intermediate phases emerge at high rates, but not at low rates (Figure 5).[11]
Li et al: NMC111 of small particle size was found to not have a phase change, whereas large particle size NMC did undergo a phase transition at C/100 rate.[12]
Clare Grey’s group (Xu et al) concluded that NMC811, does not show a phase change as the material displays “continuous shifts with no obvious splitting.”[1]
Quilty et al. found that the two phases of NMC, H1 and H2, coexist for NMC622, in all cycles – they investigated cycles 1, 2, and 100. Figure 6 shows the crystal structures of H1, H2, and H3 phases in NMC.[9] (Remember this one, it’s important).
Figure 6
Crystal structures of the (A) H1, (B) H2, and (C) H3 phases of NMC. Lithium is green, oxygen is red, and the Ni/Mn/Co is blue.[9]
Controversy
In March 2021, a Nature Materials paper came out titled “Fictitious phase separation in Li layered oxides driven by electro-autocatalysis”.[13] This paper challenges traditional interpretations of phase transitions in lithium-layered oxides. I lost sleep over this paper, not going to lie to you.
The researchers observed a very clear 003 peak splitting in NMC111, a characteristic often associated with the presence of two distinct phases, H1 and H2.
However, they argue that H1 and H2 are actually the same phase, despite the apparent peak splitting in diffraction patterns.
Their findings suggest that during slow cycling, H1 can transform into H2 without any visible peak splitting, indicating a continuous structural evolution rather than a distinct phase transition.
The authors conclude that H1 and H2 must be thermodynamically identical, meaning that the previously assumed phase separation is merely an illusion ("fictitious") rather than a real physical transition.
Figure 7
Phase evolution of NMC111 during the second cycle at different cycling currents, analyzed with operando XRD. SEM images of agglomerate particles (a) and those cycled at fast (4C) (b) and slow (C/15) (c) currents. Diffraction and electrochemical data are combined to share the capacity axes. Intensity plots of the (003) peak show bifurcation during fast delithiation, while continuous shifts occur under slow conditions. Line plots of the (003) peak at selected lithium fractions reveal non-unimodal evolution during fast delithiation, contrasting with the progressive shift under slow conditions.[13]
Fine Print and Logistics (Rate, Electrode Formulation, Cycles)
• C/15 as slow rate, 4C as fast rate
• Low active material loading (40% NMC111: 40% carbon black: 20% binder by weight)
• Thin electrode (~20μm) = better to XRD!
• One slow-forming cycle at C/20
• Operando XRD measurements on pouch cells during the second and tenth cycles
Fictitious? Bimodal existence of NMC phases or “Peak Bifurcation”
The study highlights that fictitious phase separation in lithium-layered oxides, such as NMC (LiNiₓMnᵧCo₁₋ₓ₋ᵧO₂), is a repeatable, non-equilibrium effect that results from charge redistribution rather than an actual thermodynamic phase transition. This phenomenon leads to a bimodal distribution of NMC phases, meaning that different lithium concentrations appear to coexist, even though they do not represent distinct phases.
Persistence Across Cycles and Materials
The researchers found that apparent phase separation continues in later cycles, even in NMC111 and NMC532, which are known for their structural stability.
Other research groups, including the Takeuchi group [9], have also observed similar effects in NMC622, supporting the idea that this phenomenon is not limited to a specific NMC composition.
This study challenges the conventional understanding in materials science, where phase heterogeneity is often attributed to diffusion limitations. Instead, it suggests that apparent phase separation is a dynamical artifact caused by self-reinforcing electrochemical processes.
“Electro-Autocatalysis”
NMC has a variable exchange current density as a function of lithiation state, triggering what the authors call “electro-autocatalysis”. The interfacial exchange current of NMC increases with the extent of delithiation, causing partially delithiated NMC to continue delithiating faster and further reinforcing spatial inhomogeneities.
In electrochemical systems, autocatalysis refers to a process where a reaction accelerates itself as it progresses.
In this case, the paper argues that Li intercalation and deintercalation do not occur uniformly across the electrode. Instead, small fluctuations in local lithium concentration trigger a feedback loop that amplifies these fluctuations.
This self-reinforcing mechanism causes spatially inhomogeneous charge distributions, which can mimic the effects of phase separation in diffraction data, even when no true thermodynamic phase boundary exists.
First-Cycle Effect and Its Role in Phase Separation
The study also reveals an important observation regarding the first cycle of lithiation/delithiation:
If the material is not fully lithiated at the start of charge, the extent of fictitious phase separation is significantly reduced.
This is because, during galvanostatic delithiation (constant-current extraction of Li), the C-rate to exchange current density (j₀) ratio is highest near a lithium fraction of 1.0 (fully lithiated state).
By starting the reaction from a partially reacted state, one can avoid the most unstable charge regions, thereby mitigating the fictitious phase separation effect and increasing the threshold C-rate required for its reappearance.
Appearance in Later Cycles
After the first cycle, the overpotential decreases, while the exchange current density (j₀) increases.
As a result, in later cycles, a higher C-rate is needed to observe a similar fictitious phase separation effect compared to the first cycle.
The threshold rate for triggering the effect increases, meaning that the material becomes more resistant to these non-equilibrium effects over time.
"Fictitious" Phase Separation?
Traditionally, peak splitting in X-ray diffraction (XRD) was taken as evidence of two distinct phases (H1 and H2) coexisting in a material like NMC111. However, this paper suggests that the apparent phase separation results from autocatalytic effects in lithium redistribution rather than an actual structural transformation.
During slow cycling, lithium diffusion has more time to homogenize, so the autocatalytic feedback is weaker, and no visible peak splitting occurs.
During fast cycling, small concentration fluctuations get amplified, leading to strong diffraction peak splitting—but this does not correspond to an actual phase transition.
Instead of a true thermodynamic phase separation between H1 and H2, the system self-organizes into regions with slightly different Li concentrations due to autocatalytic charge redistribution. The diffraction data reflects this dynamic inhomogeneity rather than a true two-phase coexistence.
While I would never go so far as to use the term “fictitious,” I do agree with the authors that H1, H2, and H3 are not technically unique phases: all having the same hexagonal structure with various lattice parameters. It is only the coexistence of phases and the fact that H3 tends to occur during cell failure that makes these notable. And in my continued efforts to recognize that certain verbiage is not technically correct, but continue to use it, I will keep referring to H1, H2, and H3 as “phases.” However, this nightmare-causing paper did gift us the term “peak bifurcation”, and for that, we are incredibly grateful.
Keep up the great work, Chueh group!
Sources
1. Xu, C.; Reeves, P. J.; Jacquet, Q.; Grey, C. P., Phase Behavior during Electrochemical Cycling of Ni‐Rich Cathode Materials for Li‐Ion Batteries. Advanced Energy Materials 2020, 11 (7).
2. Van der Ven, A.; Aydinol, M. K.; Ceder, G., First Principles Evidence for Stage Ordering in LixCoO2. J. Electrochem. Soc 1998, 145, 2149-2155.
3. Van der Ven, A.; Aydinol, M. K.; Ceder, G., First-principles investigation of phase stability in LixCoO2. PHYSICAL REVIEW B 1998, 58.
4. Chen, Z.; Lu, Z.; Dahn, J. R., Staging Phase Transitions in LixCoO2. Journal of The Electrochemical Society 2002, 149 (12).
5. J.P. Peres, F. W., C. Delmas, Lithium/vacancy ordering in the monoclinic LixNiO2 (0.50≤x≤0.75) solid solution. Solid State Ionics 1998.
6. X.Q. Yang , X. S., J. McBreen, New findings on the phase transitions in Li1−xNiO2 in situ synchrotron X-ray diffraction studies. 1999.
7. Croguennec, L.; Pouillerie, C.; Mansour, A. N.; Delmas, C., Structural characterisation of the highly deintercalated LixNi1.02O2 phases (with x ≤ 0.30). Journal of Materials Chemistry 2001, 11 (1), 131-141.
8. de Biasi, L.; Schiele, A.; Roca-Ayats, M.; Garcia, G.; Brezesinski, T.; Hartmann, P.; Janek, J., Phase Transformation Behavior and Stability of LiNiO2 Cathode Material for Li-Ion Batteries Obtained from In Situ Gas Analysis and Operando X-Ray Diffraction. ChemSusChem 2019, 12 (10), 2240-2250.
9. Quilty, C. D.; Bock, D. C.; Yan, S.; Takeuchi, K. J.; Takeuchi, E. S.; Marschilok, A. C., Probing Sources of Capacity Fade in LiNi0.6Mn0.2Co0.2O2 (NMC622): An Operando XRD Study of Li/NMC622 Batteries during Extended Cycling. The Journal of Physical Chemistry C 2020, 124 (15), 8119-8128.
10. Li, Z.; Chernova, N. A.; Roppolo, M.; Upreti, S.; Petersburg, C.; Alamgir, F. M.; Whittingham, M. S., Comparative Study of the Capacity and Rate Capability of LiNiyMnyCo1–2yO2 (y = 0.5, 0.45, 0.4, 0.33). Journal of The Electrochemical Society 2011, 158 (5).
11. Zhou, Y.-N.; Yue, J.-L.; Hu, E.; Li, H.; Gu, L.; Nam, K.-W.; Bak, S.-M.; Yu, X.; Liu, J.; Bai, J.; Dooryhee, E.; Fu, Z.-W.; Yang, X.-Q., High-Rate Charging Induced Intermediate Phases and Structural Changes of Layer-Structured Cathode for Lithium-Ion Batteries. Advanced Energy Materials 2016, 6 (21).
12. Li, J.; Shunmugasundaram, R.; Doig, R.; Dahn, J. R., In Situ X-ray Diffraction Study of Layered Li–Ni–Mn–Co Oxides: Effect of Particle Size and Structural Stability of Core–Shell Materials. Chemistry of Materials 2015, 28 (1), 162-171.
13. Park, J.; Zhao, H.; Kang, S. D.; Lim, K.; Chen, C. C.; Yu, Y. S.; Braatz, R. D.; Shapiro, D. A.; Hong, J.; Toney, M. F.; Bazant, M. Z.; Chueh, W. C., Fictitious phase separation in Li layered oxides driven by electro-autocatalysis. Nat Mater 2021.