Particle size distributions for cellulose nanocrystals measured by atomic force microscopy: an interlaboratory comparison
Michael Bushell,Juris Meija,Maohui Chen,Warren Batchelor,Christine Browne,Jae-Young Cho,Charles A. Clifford,Zeinab Al-Rekabi,Oriana M. Vanderfleet,Emily D. Cranston,Malcolm Lawn,Victoria A. Coleman,Gustav Nyström,Mario Arcari,Raffaele Mezzenga,Byong Chon Park,ChaeHo Shin,Lingling Ren,Tianjia Bu,Tsuguyuki Saito,Yuto Kaku,Ryan Wagner,Linda J. Johnston
DOI: https://doi.org/10.1007/s10570-020-03618-4
IF: 6.123
2021-01-06
Cellulose
Abstract:Particle size measurements of cellulose nanocrystals (CNCs) are challenging due to their broad size distribution, irregular shape and propensity to agglomerate. Particle size is one of the key parameters that must be measured for quality control purposes and to differentiate materials with different properties. We report the results of an interlaboratory comparison (ILC) which examined atomic force microscopy (AFM) data acquisition and data analysis protocols. Samples of CNCs deposited on poly-L-lysine coated mica were prepared in the pilot laboratory and sent to 10 participating laboratories including academic, government and industrial organizations with varying levels of experience with imaging CNCs. The participant data sets indicated that the central location, width and asymmetry varied considerably for both length and height distributions. To deal with this variability we used a skew normal distribution to model the data from each laboratory and to obtain the consensus distribution that describes the CNC particle size. The skew normal distribution has 3 parameters: a central location (mean), distribution width (standard deviation) and asymmetry (shape) factor. This approach gave consensus distributions with mean, standard deviation and asymmetry factor of 94.9 nm, 37.3 nm and 6.0 for length and 3.4 nm, 1.2 nm and 2.8 for height, respectively. The use of multiple probes and/or deterioration of the probe with increased use are significant contributing factors to the variability in mean length between laboratories. There is less variability in height across participating laboratories and tests of applied imaging force indicate that it is possible to image without significant compression of the CNCs. The number of CNCs necessary to obtain a reliable data set depends on the probes and operating conditions, but with careful control of various parameters analysis of 250 and 300 CNCs should provide consistent data sets for height and length, respectively for one sample. Comparison of AFM with transmission electron microscopy (TEM) data obtained in the same ILC demonstrated excellent agreement between measured lengths for the 2 methods. By contrast AFM height was approximately one half the TEM width, a result that indicates the presence of a significant number of laterally agglomerated particles, consistent with literature data.
materials science, textiles,polymer science, paper & wood