3_Gresina.indd 27Gresina, F. Hungarian Geographical Bulletin 69 (2020) (1) 27–39.DOI: 10.15201/hungeobull.69.1.3 Hungarian Geographical Bulletin 69 2020 (1) 27–39. Introduction Grain size is a fundamental property of the soils and sediments, which can provide in- formation on their origin with particular re- gard to transport dynamics, deposition and post-depositional alterations of sedimentary mineral particles. These properties can be deciphered from the particle size distribu- tion. Generally, the sedimentary deposits may contain a wide range of particle sizes from boulder fraction to the smallest size, clay and colloid (McCave, N. and Syvitski, J.P.M. 1991). In most landscape development system, the particle size distribution of the constituent sediments reflects the morpho- logical characteristics associated with the physical processes of development processes (Switzer, A.D. 2013). The importance of the particle size distribution lies in resembling the physico-chemical properties of materials (e.g., particle size) which are determined by the power and capacity of the transporting agent. Furthermore, sediments and soils have 1 Eötvös Loránd University, Faculty of Science, Department of Environmental and Landscape Geography, Pázmány Péter sétány 1/c. H-1117 Budapest, Hungary; Geographical Institute, Research Centre of Astronomy and Earth Science. Budaörsi út 45. H-1112 Budapest, Hungary. E-mail: gresina.fruzsina@csfk.mta.hu Comparison of pipette method and state of the art analytical techniques to determine granulometric properties of sediments and soils Fruzsina GRESINA1 Abstract The determination of particle size distribution is a crucial issue in various fields of earth sciences (e.g., Quaternary research, sedimentology, stratigraphy, structural geology, volcanology), environmental sciences as well as di- verse industrial applications (e.g., pharmaceuticals, cement industry). New measurement techniques developed as a result of industrial demands have also gained ground in environmental and Earth sciences research. The new techniques (especially laser diffraction) have enabled the particle characterisation in the broader size-range with a more detailed resolution. Still, they have to be compared with data obtained by classical methods. In light of the above, the primary aim of our research is to examine the methods of particle size determination critically. Excessive oversimplifications of particle size analyses routinely have used in paleo-environmental and paleo-climatological reconstructions, and other sedimentary studies, as well as insufficient knowledge of the background of the applied methods, distort the interpretation of the results. Over the past four decades, laser diffraction particle size analysers have proven to be practical tools of particle size characterisation. However, the shape of the natural sediment and soil particles are irregular and, therefore, affects the particle size distri- bution results obtained by different methods. The results of the traditional pipette method differed from laser diffraction results. The presence or absence of the pretreatments did control the differences between the two techniques. The results of Fraunhofer optical method were significantly different from Mie theory because it can detect much lower volume percentages of finer particles. Grain size results of coarse-grained samples measured by different laser diffraction devices were more comparable than the results of more clayey samples. The ratios of different sizes were changed due to the hydrochloric acid and hydrogen peroxide pretreatments. The comparison of different techniques is necessary to revaluate standards in grain size measurements which can enable the shift from conventional methods to more productive and reproducible methods. Still, light scat- tering techniques have not yet been able to displace classical methods in Earth sciences completely, in contrast to industrial applications. Keywords: grain size analysis, laser diffraction, pipette method, particle shape Gresina, F. Hungarian Geographical Bulletin 69 (2020) (1) 27–39.28 specific features that depend on their particle size distribution, e.g., hydraulic properties (porosity, mass density, water content, water retention), thermal conductivity, and specific surface area (Campbell, G.S. and Shiozawa, S. 1992; Udvardi, B. et al. 2017). Before the recent paradigm shift (Blott, S.J. and Pye, K. 2006; Újvári, G. et al. 2016; Varga, Gy. et al. 2019a) particle size distribu- tion was measured by sieving and sedimen- tation method. This can be expressed as a percentage by size class, as a fraction of total dry grains by volume or weight. The relative- ly new and fast-spreading of laser diffraction methods raises the question of how similar are the data by laser particle sizing and clas- sical techniques. Numerous studies were published to discuss the difference between the results of new particle size determination methods and conventional approaches (e.g. Syvitski, J.P.M. et al. 1991; Konert, M. and Vandenberghe, J. 1997; Beuselinck, L. et al. 1998; di Stefano, C. et al. 2010; Centeri, Cs. et al. 2015a, b). The object of these studies is generally to determine and compare the clay fraction with the results obtained by the dif- ferent methods. Simultaneously, the ques- tion may arise, where to draw the boundary of the clay fraction. The limits of the ranges may change during sample preparation (ag- gregates disintegration, removal of grain coatings by chemical pretreatment). From these studies, it can be concluded that the clay content of a sample depends on its clay mineral composition and properties, includ- ing particle shape characteristics. Materials and methods Recent soil and sediment samples were in- vestigated in the present study. Various methods analysed the particle size distri- bution of Gleysol horizon B, haplic Luvisol horizon Bt and podzolic Luvisol horizon C. Furthermore, loess and sandy loess were in- cluded in the particle size studies. Samples were collected from Hungarian locations with special attention to the representation of samples with generally different granulo- metric character (from clayey to more sandy texture). Loess and sandy loess were collected from Kőszárhegy, Podzolic Luvisol and hap- lic Luvisol samples were taken from Sopron, lastly, Gleysoil was collected from Cegléd- bercel (Figure 1). Table 1. shows the applied methods which performed on the samples. Large aggregates and rock-fragments were separated using a 2 mm sieve. Before the first set of measurements, the samples were only pretreated with sodium pyrophosphate (5%) to disintegrate aggregates into individual particles. During the next series of measure- ments, the organic matter and CaCO3 coat- ings were also removed from the test samples by using hydrogen peroxide (30%) and hy- drogen chloride (10%). Pipette method and laser diffraction were used during the study. The pipette method is based on the change in density over time calculated from suspen- sion settling time (t) and depth (z), which gives all the grains in the original concentra- tion (Vs ≤ z / t). This assumes that the particles settle independently, there is no flocculation, and the temperature is constant. The solid material content of the test sample is deter- mined by evaporation and mass measure- ment. A correction value is used to subtract the dissolved salts of the sample. The cumula- tive curve of the mass of sedimentation frac- tions can be determined by the Stokes law: Vs = Δρ . g . d2 / 18μ, where Δρ is the density difference between the liquid and the particles, d is the diameter of the particles, μ is the molecular viscosity, and g is acceleration by gravity (McCave, N. and Syvit- ski, J.P.M. 1991). The pipette method expresses the number of particle size classes by weight. During the procedure of pipette method, 25–25 g of material was weighed together with 1 litre of distilled water in the settling cylinders (5 sets). Larger particles were trapped using a 250 µm sieve. Three (Podzolic Luvisol horizon C, haplic Luvisol horizon Bt, and loess) of the five samples were used for the evaluation of the pipette method. 29Gresina, F. Hungarian Geographical Bulletin 69 (2020) (1) 27–39. Deviations from the results calculated us- ing the theoretical Stokes law (suspecting spherical particle shapes) can be expected if the particles are irregularly shaped, as most clay particles have a flat, lamellar shape. The non-spherical particles settle with their maxi- mum cross-section projection perpendicu- lar to the settling direction. Consequently, this situation increases the expected tensile strength of the particle and reduces the set- tling rate. The particle shape affects the re- sults, as an overestimation of the so-called fine fraction (di Stefano, C. et al. 2010). The pipette method has more other drawbacks: it is time-consuming, highly dependent on laboratory technique and operator error (Syvitski, J.P.M. et al. 1991). It requires a large volume of samples (at least 20–25 g) for anal- ysis. Hence, the speed of the method is not sufficient to accurately analyse large num- bers of samples (Beuselinck, L. et al. 1998). The laser diffraction particle size analysis is based on the interaction of laser light and the particles, as reflection, refraction, absorption and diffraction of light (caused by the parti- cle) result in a specific light scatter pattern depending on particle size. The angle and in- tensity of scattered light are transformed into particle size distribution by different optical theories. The traditional laser diffraction ana- lysers are based solely on the principle that particles of a given size diffract the light at a given angle. The angle of diffraction increas- es with decreasing particle size (McCave, N. and Syvitski, J.P.M. 1991). Generally, two optical models can be used to calculate particle size from the light in- tensity: Fraunhofer and Mie theories (Gee, G.W. and Or, D. 2002). Fraunhofer diffrac- tion calculates only by the angle of light dif- fraction (de Boer, G.B.J. et al. 1987), while Mie theory also takes into account the events of light absorption and refraction during parti- cle size determination (Eshel, G. et al. 2004). Both theories assume that the particles are spherical. Thus, the particle diameter ob- tained from the laser diffraction is equivalent to the sphere that gives the same diffraction as the particle (di Stefano, C. et al. 2010). The sphere is the only shape which cross- sectional diameter is constant, regardless of the angle at which it is viewed. The problem is that natural particles have different cross-sections in all directions. Thus, the cross-sectional area of a non-spher- ical particle is larger than that of a sphere having the same volume as the particle. It places this tested particle in a larger size range than can be inferred from its apparent radius. Thus, there is a shift in size distribu- tion towards coarser fractions (Eshel, G. et al. 2004). The laser diffraction gives the particle size distribution as a percentage by volume. The laser diffraction devices which originate from different manufacturers usually differ from each other. These differences based on the laser systems, the number of detectors and the measuring range (Table 2). During the refraction and absorption ad- justments, the Mie theory was applied to the Fig. 1. Map of the sampling areas Table 1. Applyed grain size analysis methods in the case of different samples Samples Measurements by pipette method laser diffraction Loess Luvisol Podzolic Luvisol Sandy loess Gleysol 5 repetiton Ø Ø 5 repetition HUNGARYHUNGARY ROMANIAROMANIA UKRAINEUKRAINE SLOVAKIASLOVAKIA POLANDPOLAND CZECH REPUBLICCZECH REPUBLIC AUSTRIAAUSTRIA CROATIACROATIA SERBIASERBIA SLOVENIASLOVENIA BOSNIA- HERZEGOVINA BOSNIA- HERZEGOVINA Sopron Ceglédbercel Kőszárhegy Gresina, F. Hungarian Geographical Bulletin 69 (2020) (1) 27–39.30 Horiba Partica LA 950 V2 with a refractive index of 1.45 and a light absorption value of 0.1 (according to the recommendations of Varga, Gy. et al. 2019b). The other device which was used during the study was the Fritsch Analysette 22 Microtec Plus. Quartz refractive index of 1.54 and light absorption value of 0.1 was selected in the measure- ments settings (additionally, the Fraunhofer settings were also applied during the Fritsch measurements). Measurements were made in wet dispersion in the case of both devices (RI of water [1.33] was applied). The shape of the particles can be charac- terised by various properties, e.g., shape, roundness, and sphericity of the irregularity (Blott, S.J. and Pye, K. 2008). The shape of a particle can be described by its three-di- mensional characteristics, which are defined by the ratio of length, thickness, and width (Sneed, E.D. and Folk, R.L. 1958). Convexity describes how closely the shape of a given particle approximates the form of a real sphere (in two dimensions, this property is called circularity). If the surface of a parti- cle has significant depressions (concavity) and protrusions (convexity), its shape can be sensed irregular (Blott, S.J. and Pye, K. 2008). The method was applied to each of the five samples in order to obtain information on the shape of the particles examined. The shape information of the tested sediments and soils was provided by the Malvern Morphologi G3-ID automated stat- ic image analyser. Recently, morphological characterization of grains is a dynamically developing method for investigating various sediments (Moss, A.J. 1966; Rogers, C.D.F. and Smalley, I.J. 1993; Varga, Gy. et al. 2018; Varga, Gy. and Roettig, C.-B. 2018; Király, Cs. et al. 2019). The number of analysed grains was not sufficient for robust statistical analysis. This problem has been successfully overcome with automatized systems (Cox, M.R. and Budhu, M. 2008). Image analysis provides direct observational data of parti- cle size, and due to the automatic measure- ment technique, a large number of particles are characterised allowing us a more robust and objective granulometric description of particles compared to manual microscopic approaches (Varga, Gy. et al. 2018). Image analysis-based measurements were organ- ised into a number-based database, which can be transformed into a volumetric data- base as well. All of the particles have their identity number (ID). The applied greyscale intensity threshold was 0–45 with 20× objec- tive. The shape parameters were determined automatically. Circularity and aspect ratio were analysed in this study as attributes of the individual particles. Aspect ratio is the ratio of width and length, and circularity pa- rameter of a particle describes the propor- tional relationship between the circumfer- ence of a circle equal to the object’s projected area and perimeter. Particle size ranges of pipette method were also used in the case of laser diffraction to compare the results of the two approaches adequately. However, during comparison of the results of the two laser diffraction de- vices, the three distribution curves (Fritsch Analysette 22 Mie and Fraunhofer models, Horiba Partica LA 950 V2 Mie theory) with the original grain size bin allocations were plotted together. For samples that only have undergone laser diffraction measurements, Table 2. Measuring properties of laser diffraction devices Manufacturer Device Measuring range, µm Number of sensors Optical method Type of laser Horiba Ltd. LA-950 Laser Particle Size Analyser 0.01–3,000 n.d. Mie 650 nm (red), 405 nm (blue) Fritsch GmbH. Analysette 22 0.01–2,100 57 Fraunhofer, Mie 532 nm (green), 850 nm (infrared) n.d. = No data. 31Gresina, F. Hungarian Geographical Bulletin 69 (2020) (1) 27–39. a different representation was used. This definition applies to sandy loess and B ho- rizon of Gleysol. The same chart shows the results of untreated and pretreated samples per device. Bar chart and connected dot chart types were used to display data. Beyond the visual evaluation of the graphs, the results were compared by performing the linear re- gression analyses (n = 13). Results Comparison of laser diffraction and pipette results For the untreated sample of loess (Figure 2, a), the <2 µm fraction can be characterised by the highest volumetric proportion compared to other ranges. This trend is valid for all four distributions. The results of the pipette method were not significantly different from the laser diffraction. For the pretreated sam- ples (Figure 2, b), the proportion of the <2 µm range decreased and the volumetric contribu- tion of silt fractions (10–20 µm and 20–50 µm), increased in the case of all four methods. Based on Figure 3 (a), and Figure 3 (b), it can be stated that the hydrochloric acid and hydrogen peroxide pretreatment resulted in a more even distribution in all three cases. Typical loess distribution was obtained with a minor deviation for the smallest fractions. In the case of untreated (Figure 3, a) elemental particles, there are some differences in the dis- tribution curves, especially in the sub-micron fraction. It is also important to note that the difference between Fraunhofer diffraction and Mie theory was also apparent, although the same instrument measured it. The Fraunhofer optical model does not show a secondary max- imum in the untreated sample, in contrast to the Mie theory. Besides, the result obtained by the Horiba instrument shows the highest sec- ondary peak for the same sample. So, the dif- ference was reflected in the results obtained by the same optical model, which was measured with different manufacturers’ equipment. In the case of the pretreated sample (Figure 3, b) the difference is reduced, partly because of the larger particle size ranges converge. The regression coefficient between the two devices was above 0.94 (R2untreated = 0.93). In the case of Luvisol, similarly to loess, pretreated (Figure 2, d and Figure 3, d) sam- ples show a more uniform distribution than the untreated ones (Figure 2, c and Figure 3, c). The ratio of the <2 µm particles was materi- ally reduced by hydrogen peroxide pretreat- ment, especially in the case of Horiba particle size analyser, where the volume percentage of the <0.5 µm fraction decreased from 12 to 0 per cent. Simultaneously, the proportion of larger particle increased, especially for 31–63 µm and 63–125 µm. The regression coeffi- cient showed a higher value for pretreated samples of laser diffraction devices (R2untreated = 0.72; R2pretreated = 0.89). The third sample, the horizon C of Podzolic Luvisol, was subjected to pretreatment com- bined with the pipette method. The compari- son was made by using two series of pipette measurements. As a result, the size of the particles was influenced by the duration of pretreatment (Figure 2, g). The mass propor- tions of the finest and the coarsest size frac- tions showed a significant decreasing trend as a function of pretreatment time. A general increase of silt-sized particles could be de- tected after longer pretreatment. The parallel measurements size distribution with longer pretreatment time was much more similar to the results obtained by laser diffraction. That is why it was used to compare the different method’s grain size distributions. In the case of Podzolic Luvisol sample, there were substantial differences between the un- treated and pretreated samples. Figure 2 (e) shows that the distributions of untreated sam- ples are following two types of curves. The Mie and Fraunhofer results of Fritsch device are almost identical. However, they differ from the Horiba values and the results of the pipette method. The former has detected a larger ratio in size range of <20 µm. The latter shows a steady increase in diameter towards the larger particles. As an effect of the pretreat- ment (Figure 2, f), the size data which were obtained by Fritsch device got closer to the Gresina, F. Hungarian Geographical Bulletin 69 (2020) (1) 27–39.32 Fig. 2. Particle size distributions of the samples as regards the results obtained by the laser diffraction method and pipette. – Pipette = percentage by weight; Laser diffraction = percentage by volume; a = Loess, untreated; b = Loess, pretreated; c = Luvisol, untreated; d = Luvisol, pretreated; e = Podzoic Luvisol, untreated; f = Podzolic Luvisol, pretreated; g = Podzolic Luvisol, pretreated by two methods <2µm 2–5µm 5–10µm 10–20µm 20–50µm 50–250µm 250µm< 0 10 20 30 40 50 60 70 80 90 100 % Grain size <2µm 2–5µm 5–10µm 10–20µm 20–50µm 50–250µm 250µm< 0 10 20 30 40 50 60 70 80 90 100 % Grain size <2µm 2–5µm 5–10µm 10–20µm 20–50µm 50–250µm 250µm< 0 10 20 30 40 50 60 70 80 90 100 % Grain size <2µm 2–5µm 5–10µm 10–20µm 20–50µm 50–250µm 250µm< 0 10 20 30 40 50 60 70 80 90 100 % Grain size <2µm 2–5µm 5–10µm 10–20µm 20–50µm 50–250µm 250µm< 0 10 20 30 40 50 60 70 80 90 100 % Grain size <2µm 2–5µm 5–10µm 10–20µm 20–50µm 50–250µm 250µm< 0 10 20 30 40 50 60 70 80 90 100 % Grain size Pretreated for a shorter period Pretreated for a longer period Pretreated for a shorter period Pretreated for a longer period Pipette Horiba, Mie Fritsch, Mie Fritsch, Fraunhofer Pipette Horiba, Mie Fritsch, Mie Fritsch, Fraunhofer <2µm 2–5µm 5–10µm 10–20µm 20–50µm 50–250µm 250µm< 0 10 20 30 40 50 60 70 80 90 100 % Grain size a c e g b d f 33Gresina, F. Hungarian Geographical Bulletin 69 (2020) (1) 27–39. Fig. 3. Particle size distributions of samples based on laser diffraction results. Comparison of different instru- ments. – a = Loess, untreated; b = Loess, pretreated; c = Luvisol, untreated; d = Luvisol, pretreated; e = Podzolic Luvisol, untreated; f = Podzolic Luvisol, pretreated Horiba’s and pipette method’s results. The value of the R2 increased to 0.93 between the two laser diffraction devices obtained by the Mie theory (R2untreated = 0.0). In the case of the Fritsch instrument (Figure 3, e), there are two secondary maxima in the Fraunhofer as well as in the Mie distributions. The maximum is at 14 µm, and the other two additional peaks are at 200 µm and 1,000 µm. In the case of the Horiba instrument, the distribution is much smoother, since the smallest value of the dis- tribution is ~3.4 µm, while the maximum is at ~100 µm. All in all, the results of the two de- vices are entirely different. The pretreatment, however, resulted in much more uniform distributions of Podzolic Luvisol samples (Figure 3, f). Similar significant differences were reported by Varga, Gy. et al. (2019b). 0 1 2 3 4 5 6 0.01 0.1 1 10 100 1000 V o lu m e ( % ) Grain size (µm) Horiba, Mie Fritsch, Mie Fritsch, Fraunhofer 0 1 2 3 4 5 6 0.01 0.1 1 10 100 1000 Grain size (µm) Horiba, Mie Fritsch, Mie Fritsch, Fraunhofer 0 1 2 3 4 5 6 0.01 0.1 1 10 100 1000 Grain size (µm) Horiba, Mie Fritsch, Mie Fritsch, Fraunhofer 0 1 2 3 4 5 6 0.01 0.1 1 10 100 1000 Grain size (µm) Horiba, Mie Fritsch, Mie Fritsch, Fraunhofer 0 1 2 3 4 5 6 0.01 0.1 1 10 100 1000 Grain size (µm) Horiba, Mie Fritsch, Mie Fritsch, Fraunhofer 0 1 2 3 4 5 6 0.01 0.1 1 10 100 1000 Grain size (µm) Horiba, Mie Fritsch, Mie Fritsch, Fraunhofer a c e b d f V o lu m e ( % ) V o lu m e ( % ) V o lu m e ( % ) V o lu m e ( % ) V o lu m e ( % ) Gresina, F. Hungarian Geographical Bulletin 69 (2020) (1) 27–39.34 Comparison of different optical models of laser diffractometry and effects of sample pretreatments Results of laser diffraction measurements of sandy loess samples are presented in Figure 4 (a) and (b). Volumetric proportions of <10 µm fractions of treated and untreated Horiba results were significantly different from each other; the pretreatment resulted in a sub- stantial decrease in this range, but, the two secondary maxima remained the same. The Mie results of Fritsch device shows a similar tendency: as a result of the pretreatment (Fig- ure 4, b), the percentage of <10 µm fraction has decreased. The value of the regression coefficient increased as a result of pretreat- ment as measured by the Horiba and Fritsch instrument using Mie theory: R2treated = 0.87; while R2untreated = 0.43. It is worth noting that results calculated by using the Fraunhofer theory were profound than the grain sizes of Mie settings. However, the three laser dif- fraction curves are moving together in the ranges above 100 µm. The result of the ho- rizon B of Gleysol obtained by the Horiba laser diffraction particle size analyser shows typical characteristics of the tendency in the literature (Figure 4, c), that the proportion of smaller particle size diameters increases after pretreatment procedures (di Stefano, C. et al. 2010). The same tendency can be ob- served in the case of Fritsch device (Figure 4, d), with only a slight shift towards smaller particle size ranges. In the case of the Fraun- hofer optical model, it should be emphasised that the pretreatment did not perform the expected result, as no secondary maximum was achieved after the pretreatment. How- Fig. 4. Particle size distributions of samples based on laser diffraction results. Comparison of different instruments. – a = Sandy loess, Horiba Partica LA 950 V2; b = Sandy loess, Fritsch Analysette 22; c = Gleysol, Horiba Partica LA 950 V2; d = Gleysol, Fritsch Analysette 22 0 2 4 6 8 10 12 0 10 20 30 40 50 60 70 80 90 100 0.01 0.1 1 10 100 1000 V o lu m e (% )V o lu m e ( % ) Grain size (µm) Horiba, Mie untreated Horiba, Mie pretreated a 0 1 2 3 4 5 6 0 10 20 30 40 50 60 70 80 90 100 0.01 0.1 1 10 100 1000 V o lu m e (% )V o lu m e ( % ) Grain size (µm) Horiba, Mie untreated Horiba, Mie pretreated c 0 2 4 5 6 7 8 0 10 20 30 40 50 60 70 80 90 100 0.01 0.1 1 10 100 1000 V o lu m e (% )V o lu m e ( % ) Grain size (µm) Fritsch, Mie untreated Fritsch, Mie pretreated b 0 1 2 3 4 0 10 20 30 40 50 60 70 80 90 100 0.01 0.1 1 10 100 1000 V o lu m e (% )V o lu m e ( % ) Grain size (µm) Fritsch, Mie untreated Fritsch, Mie pretreated d 3 1 35Gresina, F. Hungarian Geographical Bulletin 69 (2020) (1) 27–39. ever, the value of the regression coefficient decreased in this case as a result of pretreat- ment between Horiba and Fritsch instrument results: R2treated = 0.63 and R 2 untreated = 0.8. Table 3 summarises the average values of two Malvern Morphologi G3-ID particle shape parameters (circularity and aspect ra- tio). The size-dependence of these properties were also tested; the circularity values de- crease from the smaller fractions towards the larger sizes (Figure 5, a–e), but larger grains typically have high aspect ratio, although their circularity parameter is relatively low. Discussion By the spread of new techniques, several research groups have tried to compare and match the results obtained with different techniques. These studies reported the clear uncertainties of determination of the clay and fine silt fractions, and unknown particle morphology was proposed to be a possible cause of the mismatches (Konert, M. and Vandenberghe, J. 1997; Beauselinck, L. et al. 1998; Pieri, L. et al. 2006; di Stefano, C. et al. 2010). The results of this paper only partially reflect the trends found in the literature. According to di Stefano, C. et al. (2010), the laser diffraction underestimates the propor- tion of clay fraction compared to the pipette method, however, this tendency is only par- tially true for our results since the ratios of different size fractions (including the clay- sized particles) were changed (primarily) due to the hydrochloric acid and hydrogen peroxide pretreatments. In the case of loess, the laser diffraction detected a higher pro- portion of the <2 µm fraction in the untreated sample than the pipette method. However, this phenomenon reversed as a result of car- bonate removal. The smallest particles of Luvisol has practically disappeared accord- ing to pipette method after the pretreatment (<2 µm fraction of the untreated sample was ~29.8%). The case of Podzolic Luvisol was different; the emphasis was on the >10 µm ranges and their changes. These phenomena are because the investigated materials had different mineral composition and organic content. The disintegration of aggregates and removal of grain coatings which were responsible for the larger grain size these es- tablished smaller particles, even nanometres in size which can no longer be detected by these methods. So the emphasis shifted to- wards the relatively larger size ranges. The particle size distribution measured by laser diffraction particle size analyser does not match the values determined by classi- cal methods, which has several causes. Laser diffraction gives a percentage by volume, whereas conventional methods (sedimenta- tion, sieving) give a percentage by weight. The result of laser diffraction is generally independent of the density of the particles, whereas the pipette method is based on the change of density over time (Syvitski, J.P.M. et al. 1991). These few differences are enough to give different results for the same sample. The literature on the results of laser dif- fraction instruments is controversial, ac- cording to di Stefano, C. et al. (2010) there is no significant difference between the cu- mulative distribution curves determined by the two optical theories. While Baywel, L.P. and Jones, A.L. (1981) and de Boer, G.B.J. et al. (1987) report significant differences in the smaller grain size ranges. Based on our study, results of Fraunhofer diffraction are significantly different from Mie theory, because it can detect much lower volume percentages of finer particles. This theory assumes that the laser beam is parallel, and the sensors are at a great distance relative to Table 3. Comparison of the volume-weighted mean shape properties of the untreated tested samples, Malvern Morphologi G3-ID Samples Circularity Aspect ratio Loess Luvisol horizon Bt Podzolic Luvisol horizon C Sandy loess Gleysol horizon B 0.921 0.960 0.884 0.933 0.944 0.841 0.879 0.821 0.844 0.885 Gresina, F. Hungarian Geographical Bulletin 69 (2020) (1) 27–39.36 the size of the diffracted particle (Loizeau, J.L. et al. 1994). If the particles are larger than the wavelength of the light, the interaction can be interpreted as diffraction (de Boer, G.B.J. et al. 1987). The Fraunhofer theory be- comes inapplicable when the particle diam- eter approaches the wavelength of light. As the refraction of the grain falls within this size range, this principle is no longer appli- cable (Loizeau, J.L. et al. 1994). Therefore, any comparison of fine-grained (clay- and fine silt-sized) fractions measured by different laser and traditional methods will provide different results for each sedimentary sample (Varga, Gy. et al. 2019b). According to Varga, Gy. et al. (2019b), if only one laser diffraction device is used with the same optical settings for all samples from the investigated profile, the significant relative changes of measured data and cal- culated values will reveal the general trends. Nevertheless, absolute values can only be compared if the same optical settings and the same devices were used. Unfortunately, Fig. 5. Comparison of the volume weighted mean shape properties of the size ranges of the untreated tested samples, Malvern Morphologi G3-ID. – a = Loess; b = Luvisol; c = Podzolic Luvisol; d = Gleysol; e = Sandy loess 37Gresina, F. Hungarian Geographical Bulletin 69 (2020) (1) 27–39. a large proportion of research papers still do not discuss the applied laser diffraction measurement settings properly, and the specification of the applied optical approach and complex refractive index are the most commonly missing pieces of information (Varga, Gy. et al. 2019b). All in all, Horiba and Fritsch grain size results of samples with a higher volumetric proportion of larger particles (sandy loess, Podzolic Luvisol, loess) were more compa- rable than the results of more clayey sam- ples. This may be because the coarse-grained samples are characterized by a higher pro- portion of more spherical particles than the clayey Gleysol with more irregular mineral grains. The shape of sand-sized grains is more similar to a sphere than the particles of smaller size ranges (Polakowski, C. et al. 2014). Therefore, the methods were more compatible with sandy samples. The results of particle shape analysis by optical micro- scope were also contradictory since particle circularity of smaller size ranges (<4 µm) were closer to 1 than the larger grains. This was contrary to the trend found in the lit- erature. The reason for this was probably due to the presence of adhesives (CaCO3, organic matter ), which formed aggregates in the sample, thus were distorting the shape distribution. The number of pixels decreases with grain size. The smaller particles have smaller area covered by pixels, which can lead to simplified shape properties compared to larger grains. In the case of Luvisol the cir- cularity property was the highest among the other samples. This sample had a relatively large amount of small particles (0.5–2.0 µm) which did not led to proper shape proper- ties over the whole sample. Consequently, analysis by separating aggregates is war- ranted. However, the device does not show a three-dimensional image of the shape of the particles. The third dimension of particles cannot be accurately determined by automat- ed static image analysis as the orientation of individual particles is not random, they are facing into the CCD-camera with their larg- est surface area (Varga, Gy. et al. 2018). Conclusions Nowadays, the laser diffraction technique is one of the most advanced methods for deter- mining particle size distribution. In contrast to the classical techniques, these measure- ments are faster, more reproducible, and need a relatively small amount of material. Depending on the purpose of the measure- ments, the question of the need for chemical pretreatment of samples has to be taken into account too. It can be stated that it greatly influences the obtained results. It may be questionable to what extent it is advisable to use hydrochloric acid pretreatment for loess since a significant part of the sediment is composed of carbonate. Is it worth remov- ing carbonates completely if they build up real grains? If the purpose of the measurements with different devices is data harmonization, it is advisable to use the same unit of meas- urement. It is not advisable to represent the measurement methods together in various dimensions, treating them in the same plane since they measure different properties of the particles. When combining data from dif- ferent laser diffraction devices, special care should be taken to ensure that measurements are made with the same optical adjustment, with particular reference to the value of the refractive index. In the case of measurements of the smallest grain fractions, the obtained results must be treated with caution. The la- ser diffraction devices which were used in this study have different structure, the re- sults obtained by them may not be the same; dual laser systems developed by different manufacturers do not operate in the same wavelength range, however, according to Varga, Gy. et al. (2019b), the wavelength of built-in laser(s) do(es) not have an effect on the results. It may be advisable to include a device capable of measuring in the submi- cron range (photon correlation spectrosco- py). An inter-laboratory comparison could help for optimizing techniques for different sediment types as well as set new standards in particle size determination methods. Also, Gresina, F. Hungarian Geographical Bulletin 69 (2020) (1) 27–39.38 it is worth getting some information about the mineral composition and shape of the particles (optical microscopy, electron mi- croscopy), which can help to explain certain phenomena and differences between the distribution curves. When using an optical microscope, it is worth considering the ir- regularity of the particles as a function of the aspect ratio of the particles rather than the circularity parameter. Even before the expansion of the laser dif- fraction technique, the standardisation of par- ticle size distribution was difficult. For more than two decades, Syvitski, J.P.M. et al. (1991) have stated that although there are many methods for determining particle size, none in sedimentology and geomorphology can be accepted as an uncompromising procedure. In the present study, the particle size dis- tributions were compared, however, beyond these methods there are more complex analyses which can be used depending on what prop- erty is desirable (simple statistical methods, ratio-based indicators, mathematical-statistical methods). In geomorphology and sedimentol- ogy, the determination of particle size distribu- tion is rarely the ultimate goal. The objectives are to determine the evolution of surface forms, and the conditions of transport as well as the deposition of the grains. One of its key compo- nents is the determination of particle size dis- tribution (Switzer, A.D. 2013). Interpretation of granulometric (particle size and particle shape) data enables understanding and reconstruction of sedimentation environments and processes controlling surface evolution. Acknowledgements: The study is supported by the ÚNKP-19-3 New National Excellence Program of the Ministry for Innovation and Technology. Support of the National Research, Development and Innovation Office NKFIH K120620 and KH130337 are gratefully acknowledged. R E F E R E N C E S Bayvel, L.P. and Jones, A.R. 1981. Electromagnetic scat- tering and its applications. Applied Science, London, Englewood N.J. Beuselinck, L., Govers, G., Poesen, J., Degraer, G. and Froyen, L. 1998. Grain-size analysis by laser diffractometry: comparison with the sieve-pipette method. Catena 32. (3–4): 193–208. Blott, S.J. and Pye, K. 2008. Particle shape: a review and new methods of characterisation and classifica- tion. Sedimentology 55. (1): 31–63. Campbell, G.S. and Shiozawa, S. 1992. Prediction of hydraulic properties of soils using particle-size distribution and bulk density data. In Indirect methods for estimating the hydraulic properties of un- saturated soils. Eds.: van Genuchten, M.T., Leij, F.J. and Lund, L.J., Riverside, University of California, 317–328. Centeri, Cs., Jakab, G., Szabó, Sz., Farsang, A., Barta, K., Szalai, Z. and Bíró, Zs. 2015a. Comparison of particle-size analysing laboratory methods. Environmental Engineering and Management Journal 14. (5): 1125-1135. Centeri, Cs., Szalai, Z., Jakab, G., Barta, K., Farsang, A., Szabó, Sz. and Bíró, Zs. 2015b. Soil erodibility calculations based on different particle size dis- tribution measurements. Hungarian Geographical Bulletin 64. (1): 17–23. Cox, M.R. and Budhu, M. 2008. A practical approach to grain shape quantification. Engineering Geology 96. 1–16. de Boer, G.B.J., de Weerd, C., Thoenes, D. and Goossens, H.W.J. 1987. Laser diffraction spectrom- etry: Fraunhofer versus Mie scattering. Particle Characterisation 4. (1–4): 14–19. di Stefano, C., Ferro, V. and Mirabile, S. 2010. Comparison between grain-size analyses using laser diffraction and sedimentation methods. Biosystem Engineering 106. (2): 205–215. Eshel, G., Levy, G.J., Mingelgrin, U. and Singer, M.J. 2004. Critical evaluation of use of laser diffraction for particle-size distribution analyses. Soil Science Society of America 68. 736–743. Gee, G.W. and Or, D. 2002. Particle size analysis. In Soil science society of America book series, Vol. 5. Methods of soil analysis, Part 4. Physical methods. Eds.: Dane, J.H. and Topp, G.C., Madison, WI, 255–293. Király, Cs., Falus, Gy., Gresina, F., Jakab, G., Szalai, Z., and Varga, Gy. 2019. Granulometric properties of particles in Upper Miocene sandstones from thin sections, Szolnok Formation, Hungary. Hungarian Geographical Bulletin 68. (4): 341–353. Konert, M. and Vandenberghe, J. 1997. Comparison of laser grain size analysis with pipette and sieve analysis: a solution for the underestimation of the clay fraction. Sedimentology 44. 523–535. Loizeau, J.L., Arbouille, D., Santiago, S. and Vernet, J.P. 1994. Evaluation of a wide range laser diffrac- tion grain size analyser for use with sediments. Sedimentology 41. 353–361. McCave, N. and Syvitski, J.P.M. 1991. Principles and methods of geological particle size analysis. 39Gresina, F. Hungarian Geographical Bulletin 69 (2020) (1) 27–39. In Principles, methods and application of particle size analyses. Ed.: Syvitski, J.P.M., Cambridge, Cambridge University Press, 3–22. Moss, A.J. 1966. Origin, shaping and significance of quartz sand grains. Journal of the Geological Society of Australia 13. 97–136. Pieri, L., Bittelli, M. and Pisa Rossi, P. 2006. Laser diffraction, transmission electron microscopy and image analysis to evaluate a bimodal Gaussian model for particle size distribution in soils. Geoderma 135. 118–132. Polakowski, C., Sochan, A., Bieganowsky, A., Ryzak, M., Földényi, R. and Tóth, J. 2014. Influence of the sand particle shape on particle size distribution measured by laser diffraction method. International Agrophysics 28. (2): 195–200. Rogers, C.D.F. and Smalley, I.J. 1993. The shape of loess particles. Naturwissenschaften 80. 461–462. Sneed, E.D. and Folk, R.L. 1958. Pebbles in Lower Colorado River, Texas: a study in particle morpho- genesis. Journal of Geology 66. (2): 114–150. Switzer, A.D. 2013. Measuring and analysing par- ticle size in a geomorphic context. In Treatise on Geomorphology 14. Methods in geomorphology. Eds.: Shroder, J., Switzer, A.D. and Kennedy, D.M., San Diego, CA, Academic Press, 224–242. Syvitski, J.P.M., Leblanc, K.W. and Asprey, K.W. 1991. Interlaboratory, instrument calibration experiment. In Principles, methods and application of particle size analyses. Ed.: Syvitski, J.P.M., Cambridge, Cambridge University Press, 174–193. Újvári, G., Kok, J.F., Varga, Gy. and Kovács, J. 2016. The physics of wind-blown loess: Implications for grain size proxy interpretations in Quaternary pale- oclimate studies. Earth-Science Reviews 154. 247–278. Udvardi, B., Kovacs, I.J., Fancsik, T., Konya, P., Batori, M., Stercel, F., Falus, G. and Szalai, Z. 2017. Effects of particle size on the attenuated total reflection spectrum of minerals. Applied Spectroscopy 71. (6): 1157–1168. Varga, Gy. and Roettig, C.-B. 2018. Identification of Saharan dust particles in Pleistocene dune sand- paleosol sequences of Fuerteventura (Canary Islands). Hungarian Geographical Bulletin 67. (2): 121–141. Varga, Gy., Kovács, J., Szalai, Z., Cserháti, C. and Újvári, G. 2018. Granulometric characterisation of paleosols in loess series by automated static image analysis. Sedimentary Geology 370. 1–14. Varga, Gy., Újvári, G. and Kovács, J. 2019a. Inter- pretation of sedimentary (sub)populations extracted from grain size distributions of Central European loess-paleosol series. Quaternary International 502. 60–70. Varga, Gy., Gresina, F., Újvári, G., Kovács, J. and Szalai, Z. 2019b. On the reliability and compara- bility of laser diffraction grain size measurements of paleosols in loess records. Sedimentary Geology 389. 42–53. Gresina, F. Hungarian Geographical Bulletin 69 (2020) (1) 27–39.40