Conductometric process analytical technology for real-time endpoint control of graphene oxide dialysis purification
Sergey P Popov1, Vladislav A Manuilenko1, Victor V Koshlan1
1Analytical Chemistry Division, Department of Chemistry, Lomonosov Moscow State University, Lenin Hills 1-3, Moscow, 119234, Russia.
Background:
Dialysis is the standard route for purifying aqueous graphene oxide (GO) dispersions after permanganate oxidation, yet its endpoint is fixed by convention: a preset duration of one week to one month, or dialysis to constant external conductivity. Residual Mn and Ti govern the redox, colloidal, and interfacial behavior of GO, so purification is the least controlled preparation step. No physically grounded, validated criterion exists for deciding when to renew the reagent and when purification is complete.
Results:
The hourly increment of temperature-normalized dialysate conductivity, Δκ, reports the transient state of the reactor, the exchange of the external medium and the passage of the ionic release wave from the bag, not the concentration of any single leached species. We convert it into a dimensionless rule, Ki < 0.2, from conductometric quantities alone, with a concentration-normalized β correction for stepwise supply. Under isocratic and gradient supply, the rule was validated against stage-wise maxima of Mn release by inductively coupled plasma optical emission spectrometry (spike recovery 95-103%); for the more slowly released Ti it marks the onset of the release wave rather than its maximum; infrared, X-ray photoelectron and Raman spectroscopy showed removal of sulfur-containing species and partial modification of oxygen-containing groups, without major structural change. A threshold sweep over all runs supported this value; trigger-time uncertainty did not exceed 0.25 h for most stages (Monte Carlo). Criterion-driven exchange removed 76-98% of Mn, 97.8% under gradient supply, against 64.9% under static dialysis.
Significance:
Taking the hourly derivative of a single low-cost conductometric signal turns a blind purification step into a measured process with an explicit stopping rule and a stated uncertainty. Because the signal encodes reactor exchange dynamics, not the chemistry of one impurity, the criterion is formulated to be transferable to closed-loop operation and to other dialysis-purified nanomaterials; both extensions remain to be tested.
