Troubleshooting

Common Mistakes in Cathode Active Material and Precursor Manufacturing: Symptoms, Root Causes, and Fixes

Seven repeat failures in CAM and precursor plants, from hydrate basis dosing errors to undersized sampling plans, each with the symptom, root cause, and a quantified fix.

A single bad lot of NCM811 hurts. At 5 metric tons per batch and roughly 20 dollars per kilogram of finished cathode active material, one out of spec calcination run puts 100,000 dollars at risk before you count kiln time and saggar wear. Most failures trace back to a small set of repeat mistakes: wrong hydrate basis on raw materials, drifting pH probes, sloppy lithium ratio math, and moisture data nobody trusts. This guide walks through the mistakes seen most often across precursor co-precipitation, calcination, washing, and finishing, with the symptom you will notice on the floor, the root cause behind it, and a fix you can put a number on.

Mistake one is dosing metal sulfates on the wrong molecular weight basis. Nickel sulfate hexahydrate runs 262.85 g/mol against 154.75 g/mol for the anhydrous salt, a 70 percent mass difference, and cobalt and manganese sulfates carry the same trap. The symptom is a transition metal ratio that misses target by several percent even though every scale is calibrated. The fix is to dose from the certificate of analysis assay, typically 22.2 to 22.4 percent Ni for hexahydrate crystal, and verify charges with the Metal Sulfate Usage calculator before the dissolver. Confirm blended feed by ICP every shift; a 0.5 percent Ni:Co:Mn drift is enough to fail composition spec on Ni rich chemistries.

Mistake two is trusting a pH probe in hot ammoniacal sulfate liquor for more than a shift. Hydroxide precursor co-precipitation typically runs near pH 11.0 to 11.6 at 50 to 60 C, and a probe fouled with hydroxide scale drifts 0.05 to 0.1 units in 8 to 12 hours. The symptom is a widening particle size distribution and tap density sliding below 2.0 g/cm3. Calibrate against fresh buffers every 8 hours, keep a second probe as referee, and watch soluble nickel in the filtrate; the Co-Precipitation Efficiency calculator converts filtrate ICP readings into a metal loss figure. Losing 0.3 percent of nickel to the filtrate on a 10 t/day line is real money.

Mistake three is computing lithium addition on nominal purity. Lithium hydroxide monohydrate is hygroscopic and absorbs both water and CO2 in storage, so effective LiOH content can fall 1 to 2 percent within weeks of opening a super sack. The symptom is residual Li2CO3 above 0.3 percent on finished powder, or first cycle capacity missing by 2 to 4 mAh/g. Titrate every lot, set the Li/TM molar ratio deliberately, commonly 1.02 to 1.05 for Ni rich material, and price the tradeoff with the Lithium Excess Cost calculator. Every 0.01 of excess ratio, with LiOH at 12 to 15 dollars per kilogram, compounds into six figures across a few thousand tons.

Mistake four is treating the kiln setpoint as the powder temperature. In a roller hearth kiln, bed depth above roughly 40 mm and stacked saggars create gradients of 10 to 30 C between crucible wall and core, so material at an 800 C setpoint may soak at 770 C. The symptom is elevated cation mixing on XRD and unreacted Li2CO3 concentrated in the core fraction. Map the kiln with tracer thermocouples quarterly, cap bed depth, and verify oxygen supply, since Ni rich CAM generally needs an atmosphere above 90 percent O2. The Calcination Energy calculator helps balance throughput against soak time when schedule pressure tempts you to shorten the profile.

Mistake five sits at the wash and dry step, and it cuts both ways. Under washing leaves surface sulfate and lithium residues that gel the customer's slurry; over washing leaches lithium from the particle surface and costs capacity. A workable target is residual sulfate below 300 ppm and free lithium near 0.1 percent at a water to powder ratio of 1:1 to 1.5:1, which the Wash Water Load calculator sizes. Then the dryer becomes the bottleneck: pulling washed cake from 20 percent moisture down to a 200 ppm spec takes serious thermal capacity, and the Drying Capacity calculator flags undersized vacuum dryers before moisture failures reach the certificate of analysis.

Mistake six is reporting yield without a mass balance. Plants quote 95 percent yield while ignoring sieve overs and fines, magnetic separation rejects, and transfer losses, then wonder why raw material purchases keep exceeding plan. Run the Precursor Yield calculator stage by stage and reconcile it against the Particle Size Distribution Yield result from your classifier cuts; a D50 target of 10 to 12 microns with a tight span still commonly loses 2 to 5 percent to overs and fines. Those rejects are not trash. Off spec CAM and precursor carry recoverable nickel, cobalt, and lithium, and the Scrap Recovery Value calculator shows when a recycler credit of 60 to 80 percent of contained metal value beats disposal by six figures a year.

Mistake seven is a sampling plan sized for convenience instead of risk. Magnetic impurity limits sit in the low ppb range, often below 50 ppb combined Fe, Cr, Ni, and Zn, and a contamination event from a worn valve seat is localized, so three composite samples per 5 t lot will miss it. Use the Quality Sampling Load calculator to size sample counts and test frequency against lot value and defect cost; on a 100,000 dollar lot, doubling ICP and magnetic sump checks costs a few hundred dollars against a full quarantine. Finally, log every deviation with raw numbers attached. Most repeat failures in CAM plants persist because the first occurrence was closed without data.

Published 2026-07-02.