Next Article in Journal
Experimental Comparison of Fastener Implementation Approaches in Fused Deposition Modeling
Previous Article in Journal
Assessment of Cardio-Respiratory Relationship during and after Exercise in Healthy Recreative Male Subjects: A Pilot Study
 
 
Search for Articles:
Title / Keyword
Author / Affiliation / Email
Journal
Article Type
 
 
Advanced Search
 
Section
Special Issue
Volume
Issue
Number
Page
 
Journals
Applied Sciences
Volume 14
Issue 12
10.3390/app14125169
applsci-logo
Submit to this Journal Review for this Journal Propose a Special Issue
► Article Menu

Article Menu

share Share announcement Help format_quote Cite question_answer Discuss in SciProfiles thumb_up
...
Endorse textsms
...
Comment
first_page
settings
Order Article Reprints
Font Type:
Arial Georgia Verdana
Font Size:
Aa Aa Aa
Line Spacing:
Column Width:
Background:
Article

Investigation of the Geometric Characteristics of Inhalable Particles Emitted from the Process of Grinding Dental Restorations

by
Milana Ilic Micunovic
1,
Igor Budak
1,
Djordje Vukelic
1,*,
Daniela Djurovic Koprivica
2,
Milos Kuzmanovic
2,
Boris Agarski
1 and
Tatjana Puskar
2
1
Faculty of Technical Sciences, University of Novi Sad, Trg Dositeja Obradovica 6, 21000 Novi Sad, Serbia
2
Faculty of Medicine, University of Novi Sad, Hajduk Veljkova 3, 21000 Novi Sad, Serbia
*
Author to whom correspondence should be addressed.
Appl. Sci. 2024, 14(12), 5169; https://doi.org/10.3390/app14125169
Submission received: 20 May 2024 / Revised: 8 June 2024 / Accepted: 11 June 2024 / Published: 14 June 2024
(This article belongs to the Section Applied Dentistry and Oral Sciences)
Browse Figures
Versions Notes

Abstract

:
The study concerns measurement and evaluation of the geometric characteristics of the inhalable fraction of particulate matter generated in the process of grinding dental restorations, which is a process that generates a large quantity of particulates. The research is based on measurements taken with a personal sampler, whereby the characteristics of particulates are determined based on the particle mass collected on filters. The collected filters were examined using scanning electron microscopy (SEM), and the resulting micrographs were processed through image analysis. The complex geometry of particles was examined through the analysis of 14 parameters, of which 6 define size and 8 describe morphological characteristics. Two software programs were used for the image analysis process to gather a wide range of parameters describing particle geometry. The relationship and dependence between the geometric parameters used to describe particle shape and size were investigated using multivariate analysis methods. Through correlation analysis, principal component analysis, and cluster analysis, parameter selection and reduction were performed to provide an understanding of the particles inhaled by exposed workers, which also influences the biological response of exposed organisms and the burden on the working environment.

1. Introduction

In analysing particulate matter in the environment and workplace, it is of great importance to determine the concentration of particles, as well as the duration of exposure to these concentrations, in order to obtain a more realistic assessment of the health risks to the exposed individuals. Additionally, it is important to consider particle concentrations within different size fractions [1,2,3]. Among the basic parameters that directly affect the level of influence on humans and technical systems are the shape, size, and dimensions of particles, alongside the chemical composition of particulate matter. The behaviour, deposition, and fate of each particle after entering the human respiratory system depend on their natures and the specified parameters [4,5,6]. Based on this, we can conclude that the importance of measuring particle size and shape distribution lies in the fact that the properties of dispersed materials depend on the uniformity of their distribution. Similarly, significant differences in particle size and shape play a crucial role in assessing exposure and the intensity of the organism’s biological response [7,8]. Epidemiological studies, development of standards in the field, as well as measurement and protection devices, are also based on the examination of particle size and morphological characteristics.
One of the primary interests of researchers is examining and understanding the adverse health outcomes that can result from intermittent or continuous exposure to particulate matter. These include pneumoconiosis, occupational asthma, chronic bronchitis, cancer, and systemic poisoning, such as lead poisoning, especially with prolonged exposures [9,10,11]. Additionally, there is growing interest in other diseases associated with dust inhalation, such as hypersensitivity pneumonitis and irritations, as well as a range of non-respiratory diseases that can occur even with short-term exposures [12]. Subfractions of inhalable dust include the nasopharyngeal, tracheobronchial (or thoracic), and alveolar (or respirable) fractions. The nasopharyngeal fraction includes a spectrum of particles that are deposited in the airways between the point of air entry, i.e., the mouth or nose, and the larynx. Particles smaller than 2.5 µm are deposited below the larynx before reaching the deepest parts of the lungs. These particles can later be eliminated through mucociliary clearance. They are generally finer than the nasopharyngeal fractions, and their primary deposition mechanisms are impaction and, to some extent, gravity. Smaller particles can penetrate the alveolar region, where inhaled air can reach the blood vessels. Regarding aerodynamic diameter, only about 1% of 10 μm particles reach the alveolar region, so an aerodynamic diameter of 10 μm is usually considered the upper size limit for penetration into this region. In this area, particles with an aerodynamic diameter of 2 μm are the most frequently deposited. For smaller particles, most deposition mechanisms become less effective, so deposition of particles with an aerodynamic diameter smaller than 2 μm is less efficient. For particles with an aerodynamic diameter of 0.5 μm, deposition efficiency is only 10–15% [13,14]. Most of these particles are exhaled without being deposited. However, in the case of even smaller particles, diffusion becomes a more efficient mechanism, and the likelihood of deposition in the body increases, although electrostatic forces can also play a significant role in the deposition of these fractions [11,15].
Although, in medical literature, exposure to dust is commonly associated with miners, industrial and construction workers, the materials used in dental technology have rendered dental technicians a population segment at greater risk [16,17,18]. Dentists and dental technicians are exposed to aerosols generated from different materials for the entire duration of their working hours, generated from different materials [19,20,21,22]. In scientific and professional literature, these materials are associated with chronic diseases such as respiratory system diseases, dermatological conditions, and allergies [23,24,25,26]. If dental technicians do not use appropriate protection, metal shavings, ceramic dust, acrylics, and other impurities, which are present during the production of prostheses, can harm their health. Dust particles from ceramics and carbides, silica particles, and heavy metal particles, which are also present in the process of prosthesis fabrication, can lead to pneumoconiosis, silicosis, and other pathological changes in the respiratory tract [27,28]. The underlying cause of these diseases is pulmonary fibrosis, which results from the binding and retention of silicate particles and other compounds in lung tissue. Fibrosis occurs as a non-specific reaction of the body to dust. The significance of these diseases lies in their frequent complications, such as tuberculosis, cardiorespiratory insufficiency, and acute lung infections [29,30]. Certain studies [31,32] have shown that aerosols composed of particles of different materials contain between 54 and 70% of the inhalable fraction, or particles that can reach the lungs through inhalation.
Previous research has shown that not only particle size and composition but also variations in their morphology—such as the shape and surface structure of particles—in inhalation formulations are important for the degree of penetration to the lungs. Accordingly, particles with different aerodynamic diameters can reach and be deposited in different regions of the lungs [33,34], and ultimately the physical and chemical properties of inhaled particles may affect their impact on lung physiology [35,36]. Previous studies have demonstrated that elongated particles are more likely to reach the lower airways [37], indicating that controlling particle shape may increase deposition in the airways. The question is which equipment can protect dental personnel from the effects of particles and what type of protective equipment can provide a sufficient level of protection [38,39]. For this purpose, a proper dust characterisation is necessary, and this includes the analysis of dust as a whole and the analysis of individual dust particles. The main characteristics of dust particles include size, shape and compositional homogeneity.
A number of analytical methods have been developed for particle characterisation. This has contributed to the fact that, in various scientific and engineering fields (such as colloidal chemistry and nano-materials, control and optimisation of the production of drugs, raw materials, food, and other products) there is a need for a fast and accurate method for the characterisation of particles. According to WHO statistics, approximately 2 million people die annually due to the presence of particulate matter in the workplace, with the majority of deaths caused by respiratory diseases [40,41]. Professional exposure to particles generated during material processing, particularly grinding processes, is one of the largest sources of particulate matter exposure for workers and must be monitored and characterised adequately [42,43,44].
The method of image analysis (IA), based on micrographs from optical microscopes and especially from a scanning electron microscope, is a preferable tool in the field of characterisation of powdery substances [45,46,47]. Differences in particle characterisation results are largely associated with irregular particle shapes. Although one of the limitations of this method is the fact that it is a time-consuming process, application of SEM in this area has enabled the use of image analysis methods even in the analysis of particles below the detection limit of an optical microscope. Thanks to their increased contrast quality, micrographs have also contributed to the improvement of the accuracy of IA. The quantitative characterisation of particles based on SEM micrographs has the following limitations [48]: the slowness of the particle sampling process; the large amount of data; the diversity and complexity of particle shapes; the problem of choosing form factors suitable for modelling; and the choice of method for image analysis. Nevertheless, despite the mentioned limitations, the method of image analysis (in the format of shades of grey colours) is one of the more precise methods for analysing particulate matter characteristics.
Some authors believe that particle shape can be adequately described using only one or two shape parameters [49,50]. However, others argue that particles present in the environment and workplace settings are predominantly non-isometric particles of irregular shapes, requiring a larger number of shape parameters for characterisation [51,52,53,54]. It is important to note that a common characteristic of most shape factors is their dependence on the method used in image analysis for assessing the basic dimensions of particles, especially perimeter [55].
Unlike previous studies, the goal of this research is to evaluate the results of measuring the geometric characteristics of powder materials during the grinding process of dental restorations. This research is based on the assumption that a spherical shape does not sufficiently describe the irregular geometry of the collected particles and that determining geometric characteristics and selecting geometric parameters that describe the dust particles are fundamentally important for sample characterization. The process of particle analysis and characterization involved acquiring data on geometric and surface characteristics from 2D micrographs obtained using SEM, conducting image analysis with two software programs, reducing and selecting key geometric parameters, and classifying particles based on their geometric characteristics. The relationships and dependencies between the geometric parameters used to describe particle shape and size were investigated using multivariate analysis methods. Through correlation analysis, principal component analysis (PCA), and cluster analysis, parameter selection and reduction were performed. This research contributes to the examination of micron and submicron particles, particularly the inhalable fraction of powdery substances, where the composition, size, and shape of the particles, as well as the combination of these parameters, affect the severity of the organism’s biological reaction. Additionally, it provides a more realistic assessment of the degree of pollution in the working environment and the health risks for exposed individuals.

2. Methodology

The research methodology is depicted in Figure 1. The methodology for investigating the geometric characteristics of inhalable particles emitted from the process of grinding dental restorations consists of five steps:
  • Grinding of dental restorations;
  • Sampling of emitted particles;
  • Investigation—scanning electron microscope;
  • Image analysis;
  • Statistical processing of results.
During the grinding process, air sampling for inhalable particle fractions was conducted. Personal sampling simulates the collection of particles that would be inhaled by a technician, who wears the sampling device.
For particle sampling, the EGO PLUS TT (Zambelli, Milano, Italy) personal sampler with a display setting, with conical nozzle, and a filter made of a mixture of cellulose esters was used. The basic technical specifications of the sampler and filter were as follows: flow rate 3.5 L/min, sampled volume 420 L, normalised volume 407.9 L, filter diameter 25 mm, effective filter diameter 22 mm, effective filter area 380 mm2, filter weight 0.022 g, and sampling time 30 min.
The Jeol JSM—5610LV (Jeol, Tokyo, Japan) scanning electron microscope with secondary electron image (SEI) and backscattered electron image (BEI) detectors and an EDS analyser (Jeol, Tokyo, Japan) was used. As part of the SEM analysis, sample conductivity preparation, elemental analysis (refractive index adjustment, contrast adjustment, particle classification), magnification adjustment, SEI, and BEI micrographs were carried out. In the first step, a quarter of the filter was cut with the collected particle mass, and the conductivity was prepared for analysis by SEM. Non-conductive samples required special preparation. Using the method of SEM, two types of micrographs, BEI and SEI, were obtained. BEI micrographs provided information about the topography of the sample. SEI micrographs were used to record surface irregularities (roughness). SEM was used to obtain a magnified image of an object by diffraction of high-energy electrons. The use of SEM in combination with the method of image analysis enabled a precise examination of the size and morphology of the particles, as well as the observation of the surface structure of the particles.
As part of the image analysis, the micrographs obtained on the SEM were imported and pre-processed, and particle selection and particle image extraction were carried out. For image analysis, in order to obtain particle size distribution and their geometric characteristics from 2D SEM micrographs, it was necessary to implement certain steps in order to enable proper characterisation of the samples. In a specific investigation, it is necessary to create a set of specific steps to address various image processing stages and determine the workflow. Images based on different shades of grey are converted into binary images by properly setting threshold values. Using the same threshold setting for all particles can lead to an increase in the so-called ‘halo effect’ around the particles and thus underestimate their dimensions. Therefore, a separate threshold adjustment procedure was examined and implemented for each particle. Subsequently, particles were separated from micrographs, and parameters determining their size and morphological characteristics were examined. Determination of particle geometry was conducted with the aim of their quantitative characterisation. The output from this phase was particle size parameters (Table 1) in accordance with standards [56,57,58].
Given the large number of obtained numerical parameters characterising particle geometry and the non-linear relationships between these parameters, the final step required statistical data processing. Correlation analysis, principal component analysis (PCA), and cluster analysis (CA) were used. Correlation analysis revealed the mutual connection between two variables, and the correlation coefficient was a measure of the connection between them. Principal component analysis enabled a more detailed examination of the complex relationships between parameters, dividing them into several factors, based on their characteristics. PCA allowed the original set of variables to be transformed into a new set of reduced variables and the selection of a representative geometric parameter within a group of parameters with a high degree of correlation. The principal component analysis method was applied to identify geometric characteristics on the basis of which factors were extracted. The geometric characteristics of the inhalable particles fraction were classified using a clustering algorithm, and, after that, representative parameters characterising the powdery substances (emitted particles) were extracted.

2.1. Sampling

A sample group was collected within the working environment of the Department for Prosthodontics with Dental Laboratory at the Dental Clinic of Vojvodina, Faculty of Medicine in Novi Sad. The process of grinding dental restorations was conducted on the EWL KaVo micromotor (KaVo, Biberach, Germany), with speeds ranging from 1000 to 50,000 rpm. Measurements were taken within the first zone, i.e., the personal sampler was positioned on the upper part of the technician’s chest in the breathing zone (Figure 2). The research was conducted under controlled microclimatic conditions: air flow rate of 0 m/s, temperature of 23 ± 1 °C, and relative air humidity of 35 ± 1%.
Filters made of mixed cellulose esters were used, suitable for light microscopes and SEM. Due to their texture, these filters allow for lower static particle repulsion and are less susceptible to weight changes due to moisture absorption.
In the analysis of powders of heterogeneous and, conditionally speaking, unknown composition, research is mainly based on examining the composition and concentration at a given moment. Collection and examination of the inhalable fraction of particles based on filter sampling using time-integrated methods, followed by SEM sample analysis, are rarely applied, among other reasons due to adjustment issues during analysis on the scanning electron microscope. Primarily, it is necessary to determine the optimal sampling time so that the layer of dust is not too thick. Additionally, dust is non-conductive and requires special preparation. The filter is coated with a thin layer of gold, making the sample more conductive, and enabling better micrographs to be produced. The problem with analysing dust using this method is that the particles have different compositions, so in BEI micrographs, some particles appear darker (derived from elements with lower Z), while others appear lighter (derived from elements with higher Z).
The micrographs of the examined samples collected during SEM analysis (Figure 3) were processed using the image analysis software ImageJ (v1.53) and JMicroVision (1.3.2.). Within the research, 14 geometric parameters of the particles were examined, including 6 related to their size and 8 related to particle shape. To determine the size and shape of the particles, more than 1500 particles were examined.

2.2. Image Analysis

2.2.1. ImageJ Software

In the ImageJ software, the following parameters were analysed: area, perimeter, Feret’s diameter (minimum and maximum), circularity, aspect ratio, solidity, and roundness. Regions of interest were defined on each image by excluding information that could affect the contrast or particle analysis. For each image, threshold values were set for the selected regions of interest, background intensities were defined, background noise was subtracted, and irrelevant particle pixels were discarded.
A semi-automated method was adopted within the research to investigate a large number of particles deposited on the sampler filters. After defining (framing) the desired area for particle examination, they were discarded. Following the selection of the examination zone, threshold adjustment for each micrograph was applied individually. Determining the background intensity and subtracting background noise were necessary steps. Any individual particle pixel that was not of interest was discarded.
To subtract noise and refine the edges of particles, a Gaussian blur filter was used. Subtracting background noise using the Gaussian filter involved creating a duplicate of the image and applying the Gaussian filter to it, adjusting the radius value accordingly. This filtering reduced the image to 2D convolution operations. The operation assigns greater importance to pixels on the edge compared to those towards the centre of the image, as well as to pixels at the corners of edges compared to those that are not. Therefore, when attempting to sharpen the image with a high blur radius value, the output will be an image with dominant edge pixels. After applying the Gaussian filter to the duplicate image, this image needed to be subtracted from the original (Figure 4a). The resulting image displayed particles clearly after adjusting brightness and contrast parameters (Figure 4b). Subsequently, it was necessary to determine the threshold intensity. By adjusting the threshold values, particles were shown in red, while the background was in shades of grey (Figure 4c). After confirmation, the background became white, leaving only the black particles. Each particle could be individually sharpened and the cracks resulting from contrast differences removed. If discrepancies compared to the original image were observed, particles could also be completely removed from the image after threshold adjustment.
The next step involved adjusting measurement parameters followed by particle analysis (Figure 5). The obtained results could be exported to Microsoft Office Excel 2013.
Within the particle analysis setup menu, options are available for displaying particle outlines, results, inclusion of holes, and rejection of edges (i.e., particles touching the image edge). As a result, quantitative parameters defining particle size and shape were obtained.

2.2.2. JMicroVision Software

In order to obtain more information about particle geometric characteristics and select representative parameters, SEM micrograph examination was also conducted using JMicroVision image processing software.
Similar to the previous software, spatial calibration was first performed based on a known length from the SEM micrograph, entered into the calibration menu as a known distance (in micrometres).
Next, object extraction was performed (General description > Object extraction > Colour or grey intensity threshold). Firstly, the area of the image containing particles was selected to prevent accompanying information from affecting the results. The ‘all edges’ option allowed the software to exclude particles located at the edges from the calculation, in case they were not fully captured in the micrograph. Then, the threshold intensity level was adjusted (Figure 6a), until a satisfactory level of selection was achieved (using the cursor to add colour to the particle with the ‘Add’ option (Figure 6b). Each particle could be removed from the results if any deviations were observed compared to the original image.
JMicroVision, through the Image Factory option, provides the capability of classic, non-linear, and gradient filtering. Classic or linear filtering, also known as a mathematical convolution operation, involves multiplying a group of pixels in the input image by a set of pixels in the convolution kernel. The output value produced in the spatial convolution operation is the weighted average of each input pixel and its neighbouring pixels in the convolution kernel. Linear filtering encompasses a wide range of filters for fine processing, sharpening, noise reduction, and edge filters (uniform, triangular, Gaussian, Wiener, etc.). Non-linear filtering is based on the logical separation of filters into a series of relatively simple operations (median, minimum, maximum, Kuwahara filter). Gradient filtering refers to edge detection by computing the magnitude of the gradient vector of the image in two orthogonal directions. After setting all necessary parameters in the software, we took into account quantitative data defining particle size—equivalent circle diameter, length and width—as well as parameters defining particle shape—convexity, elongation, compactness and rectangularity (Figure 7).
All quantitative parameters describing particle geometry obtained as a result of using the aforementioned two pieces of image analysis software are available in the Supplementary Materials.

2.3. Selection and Reduction Process for Geometric Parameters

2.3.1. Normalisation and Standardisation of Data

In data sets, variable values often differ in magnitude and/or units. To ensure that all terms are dimensionless and that all significant deviations are minimised, data must be standardised. Prior to principal component analysis and clustering analysis, the z-score standardisation technique was applied. Although some procedures do not require data to conform to a normal distribution, for better results, the obtained values were transformed using the Johnson transformation to avoid issues due to potential variability in the original data set.
While principal component analysis and clustering analysis are used descriptively as suitable means of summarising relationships in a large set of observed variables, assumptions about the distribution of variables are not in force. If the variables are normally distributed, the solution improves. To the extent that normality is not established, the solution degrades but can still be meaningful.
When it comes to measuring various parameters and their characteristics, data values can be highly disproportionate. In such cases, it is appropriate to apply some form of transformation before bringing them to the same scale [51,59]. Probability plot testing for normal distribution (Figure 8) was performed for all investigated particle geometry parameters, assessing whether the data fit into a specific distribution.
This option primarily created a cumulative distribution function by positioning each value in relation to the estimated cumulative probability. The scale was transformed so that the corresponding distribution formed a straight line. It displayed a 95% confidence interval. The results showed an estimation of the parameter distribution along with Anderson–Darling statistics and p-value. Since not all parameters were found to follow a normal distribution, a Johnson transformation was performed to normalise them.

2.3.2. Correlation Analysis

The correlation matrix containing the linear correlation coefficients of each pair of variables is the basis for conducting principal component analysis. One of the prerequisites for conducting PCA analysis is the relationship between the original variables, and the basis for observing a set of connected variables is the correlation matrix. To determine the degree of connection between geometric parameters describing particles, Pearson correlation coefficients were primarily calculated. The Pearson correlation coefficient indicates whether there is a linear relationship between variables. A correlation coefficient value from 0 to 1 indicates positive correlation, signifying a simultaneous increase in the values of both data sets. A correlation coefficient value from 0 to −1 indicates negative correlation, meaning that an increase in the value of one variable corresponds to a decrease in the value of the other variable. Complete correlations, i.e., correlation coefficient values of r = ±1, are not characteristic of natural systems and are most commonly associated with theoretical models. When the correlation coefficient has a value of 0, it signifies the absence of a linear relationship, indicating that knowing the values of one variable does not allow us to infer anything about the values of the other.
Correlation analysis was conducted to assess the relationship between the 14 geometric numerical parameters listed in Table 1. Parameters with a high degree of correlation were grouped together as highly correlated parameters describing similar or correlated particle characteristics. It was necessary to select only one parameter as a representative for each group. Numerical parameters with low correlation coefficients with any other parameters were identified as independent representative parameters.

2.3.3. PCA and CA Analysis

In image analysis software packages, dimensionality is a common issue that can also be a factor in degrading the performance of a given algorithm as the number of features increases. Therefore, it is often necessary to use models that perform dimensionality reduction for large-dimensional data sets. Dimensionality reduction techniques are frequently employed as a step in data processing to reduce the complexity of the data image. This allows multidimensional data to be represented by a more appropriate lower-dimensional representation. Principal Component Analysis is an effective technique for identifying and reducing data in the image analysis process and pattern recognition among the data [60,61]. PCA provides direct insight into the interrelationships between variables and offers empirical support for addressing conceptual issues regarding the underlying data structure. The method focuses on reducing linear projections of multivariate high-dimensional data to a smaller number, while retaining as much information as possible.
Principal Component Analysis was performed to select representative parameters from the group of correlated parameters. During the correlation analysis, parameters that were correlated were identified and grouped together, but the representative parameter was not determined. For this purpose, the component matrix was used to extract the parameter with the highest loading factor.
The examined geometric parameters are continuous variables; thus, for determining the minimum number of parameters needed to characterise a sample, cluster analysis can be utilised. The applied hierarchical agglomerative clustering algorithm first groups the most connected variables, then gradually reduces the number of clusters at each hierarchical level from n clusters of size 1 to a single cluster that connects all attributes (referred to as the ‘tree structure’). Clustering was performed using the method of variable grouping with the applied absolute correlation distance measure, and average linkage was selected as the linkage method. The average linkage method considers information from all pairs between two clusters, making it generally preferred. The strength of the relationship, i.e., the degree of correlation in considering the distance, was more important than the sign, hence the absolute correlation method was employed.
Correlation analysis, principal component analysis and cluster analysis were conducted using the software package Minitab 16.

3. Results

Reduction of Particle Geometry Parameters Using Correlation, PCA and CA Analysis

Correlation analysis is applied to assess the degree of association between two variables within a group of investigated parameters. It was conducted on 14 geometric parameters (6 describing size and 8 morphology of the particle). All correlation coefficients at the significance threshold of 0.05 (p < 0.05) indicate exceptional associations among variables (Table 2).
When determining the number of components for principal component analysis, the criterion of the latent root was taken into account, according to which only those factors with eigenvalues greater than 1 are considered. Based on this criterion, three components should be considered, which account for 94.2% of the total variance (Table 3). The scree test—graph (Figure 9) looks for the point where the line abruptly changes direction, and up to that point, the number of components to be included in the analysis is counted. By applying this test to determine the number of components, one additional component could potentially be included. However, considering that the eigenvalue was very low and explained only 2.3% of the total variance, and based on the criterion of the latent root, it can be seen that the first three components were optimal for defining the sample using this method. The eigenvalue indicated the proportion of total variance associated with a particular principal component. The proportion of the first principal component in the total variance was 63.8%, defining the variation in data due to the first principal component. The other components accounted for 19.5% and 10.9%, respectively. It was observed that the proportion of components gradually decreased. The first component was, of course, far more important and influential than the others.
With statistical significance of component loading at a 95% probability level, for a sample consisting of 1517 particles and an equal amount of quantitative data per parameter, a level of 0.3 loading was considered significant in interpreting the results. From the geometric component matrix (Table 4), the highest positive loadings of the first component regarding parameters defining particle size were 0.318 (L) and 0.308 D ¯ F ), with a negative loading of −0.318 (P). High loadings, if positive, indicate what the component is, and, if negative, what it is not. It can also be observed that the other size parameters in component PC1 did not deviate significantly in the results. In the second component, the highest loading was 0.302 (W), while in the third component, no parameter regarding particle size stood out.
When considering parameters defining particle morphology, the first component singled out compactness with a loading of 0.304. However, the loadings of other parameters on this component had close values. PC2 and PC3 again identified multiple parameters with loadings: 0.468 (El), 0.461 (AR), and 0.444 (RD) in the second component and 0.317 (RT) and 0.320 (AR) as positive loadings, and −0.381 (CVX) and −0.466 (SLD) as negative loadings in the third component. The relationship between the first two components is depicted in the biplot (Figure 10).
Loadings and the distribution display of parameters should greatly assist in defining the components. However, the problem lay in having too many variables (14 variables and 1517 samples), making interpretation complicated. For these reasons, it was necessary to redistribute these loadings to achieve an interpretation that made sense for all components. This was accomplished by rotating the axes in the coordinate system representing the components around the set of original data. In this particular case, three components were rotated using orthogonal varimax rotation, which maximised the sum of variances of squared factor loadings. Table 5 presents the factor loadings for each variable after rotation.
The results of the cluster analysis of particle geometric parameters are illustrated by the dendrogram in Figure 11. A shorter distance between clusters indicates a stronger correlation between variables. Size parameters remained grouped in one cluster, demonstrating the strongest correlation among these sizes. The second cluster consists of parameters such as circularity, compactness, convexity, and rectangularity. The third cluster identifies parameters such as roundness, elongation, and aspect ratio, defining the particle form, and they are separated into a distinct group, similar to the PCA analysis. The fourth cluster singles out solidity as a distinct parameter defining the morphological roughness of the particle.

4. Discussion

Due to the significance of particle characteristics in the development of models explaining penetration depth and particle behaviour after entering the human body, especially from the material grinding process as one of the major sources of particles, significant attention has been devoted to the research of particle size and shape. In image analysis software packages, dimensionality is a common problem that can also be a factor in degrading the performance of a given algorithm with an increase in the number of features. Therefore, it is generally necessary to use models that reduce dimensions for large dimensional data sets. Dimension reduction techniques are often used as a step in data processing to reduce the complexity of the data picture. The particle analysis process collected from two types of grinding was carried out through several steps: (a) collection of particles with a personal sampler that analysed the amount of inhaled air during working hours; (b) acquisition of 2D micrographs using a scanning electron microscope; (c) characterisation of particles by numerical parameters using two image analysis software; and (d) reduction and selection of representative parameters characterising particles from the process of grinding dental restorations.
The correlation analysis revealed a high degree of correlation between parameters describing particle size. The correlations among these parameters were positive, indicating a harmonious growth of values within this data set. Particle size parameters did not show a high degree of correlation with particle shape parameters. Based on numerous Pearson correlation coefficient values, a certain degree of correlation was observed between parameters describing particle morphology and the equivalent circle diameter, perimeter, width, length, and mean Feret’s diameter of the particle. Additionally, it is evident that elongation, convexity, circularity, roundness, solidity and compactness were negatively correlated with size parameters. On the other hand, aspect ratio and rectangularity were positively correlated with size parameters. Given that size parameters were positively correlated with each other, the uniformity regarding exclusively negative or positive correlations with shape parameters is understandable. Elongation exhibited a significant positive association with roundness and a negative correlation with aspect ratio. Convexity showed a remarkable positive correlation with circularity and compactness, as well as circularity with compactness, while there was a high negative correlation between rectangularity and compactness, as well as between aspect ratio and roundness. A strong association existed between circularity and solidity, and in the negative direction between rectangularity and convexity, circularity and solidity. A moderate positive correlation existed between convexity and solidity, circularity and roundness, and solidity and compactness.
During PCA analysis, the loadings of parameters describing particle size on the first factor were lower compared to the loadings of the same parameters on the first component before rotation. Now, it was easier to determine what Factor 1 actually represented and what it did not. The factor showed positive loadings of 0.575 (El), 0.579 (AR), and 0.533 (RD). Based on the extracted parameters, it can be seen that the first factor defined particle form. The second factor isolated all size parameters. Positive loadings included Feret’s diameter (0.397) and particle length (0.372), while negative loadings included area (−0.429), perimeter (−0.362), width (−0.4), and equivalent circle diameter (−0.431). Negative loadings indicated what the factor did not represent. Due to the way rotation was performed, the factor was named based on the highest loading regardless of the sign. The highest factor loading actually indicated the variables that had the strongest correlation with that factor. The third factor isolated parameters defining particle morphology, specifically those on which F1 did not show loadings. Positive loadings were observed on the rectangularity parameter (0.407), while parameters CVX (−0.460), Circ (−0.426), SLD (−0.520), and CP (−0.393) exhibited negative loadings. Based on the obtained results, each factor could be identified, so Factor 1 represents form, Factor 2 size, and Factor 3 particle roughness.
The results obtained using the clustering method enabled the extraction of parameters defining the sample itself. Considering the PCA results, correlation analysis, and clustering analysis together, it can be concluded that the equivalent circle diameter emerged as the parameter that best described the particle size in this sample. The correlation analysis showed a strong association with other size parameters. PCA analysis exhibited the highest loading on this parameter, as well as in clustering, where DCE showed significant correlation with other variables, thus being singled out as representative. This result is consistent with the relationships observed in the correlation matrix and also supports the results of the PCA. The results also allowed the reduction of eight shape parameters to the three defining form, morphological, and textural roughness. El, RD, and AR were singled out into a separate group, where AR exhibited slightly higher value on the first factor in PCA analysis (Table 5) and was thus chosen as the form representative parameter. For morphological roughness, the solidity parameter stood out independently in the clustering analysis and also showed the highest loading in the third factor of PCA analysis. Convexity isolated textural roughness with the highest loading in the third factor from this group of data. Solidity and convexity were also singled out in correlation analysis as independent parameters, with weaker correlation with other parameters.
To reduce the number of variables, considering the results of multivariate statistical analysis on the sample generated by the dental restoration grinding process, one size parameter and three shape parameters that sufficiently defined particle geometry in the sample can be singled out. Area and equivalent circle diameter emerged with almost equal loadings and high correlation with other size parameters. Area showed slightly higher value in PCA analysis, allowing it to be singled out as representative. Regarding particle shape descriptors, a correlation between solidity and circularity could already be observed in the correlation analysis. Circularity showed significantly higher loading in the F3 principal component analysis, while, in the clustering analysis, the loadings and associations were at the same level. Based on the results, circularity emerged as a representative of particle morphological roughness. Convexity was singled out in correlation analysis as an independent parameter with lower association with other shape parameters, and the highest positive loading in the second factor of PCA analysis, hence representing textural roughness. Among the particle form descriptors, aspect ratio emerged with the highest loading.
According to the authors’ knowledge, this is the first study examining the characteristics of inhalable particles collected from the grinding of dental restorations. Additionally, a large number of parameters examining particle characteristics were investigated, collected by image analysis using two commercial software programs. Quantitative data were analysed and compared. This methodology allows application to other samples for characterising particle geometric characteristics, as well as to particles from the respiratory fraction. Furthermore, it provides a basis for determining the optimal surface texture of particles generated by grinding. Moreover, assessments of the impact of particle surface textures on their pathophysiological responses should be further explored, especially in the case of ultrafine/nanoscale particles.

5. Conclusions

It has been determined that the combined effects of particle size and shape influence the assessment of exposure and the intensity of the organism’s biological response. This study identified four groups describing grinding particles: particle size, form, morphological, and textural roughness parameters.
By analysing the results obtained from the grinding process, we could determine that out of the six parameters defining particle size, the equivalent circle diameter emerged as the representative. Aspect ratio was identified as the key parameter for particle form, and convexity for textural roughness. In the sample from the dental laboratory, solidity was independently identified as the representative of morphological roughness. Based on numerous values of these parameters, it has been concluded that the sample is characterised by fine particles, with partially expressed symmetry, but low textural and morphological roughness.
Future research will be focused on the analysis of a larger number of output parameters (composition of the particles, cytotoxicity, etc.), as well as on modelling, multi-criteria analysis and optimization of the process of grinding dental restorations.

Supplementary Materials

The following supporting information can be downloaded at: https://www.mdpi.com/article/10.3390/app14125169/s1.

Author Contributions

Conceptualisation, M.I.M. and I.B.; methodology, M.I.M. and I.B.; software, B.A. and D.V.; validation, M.K. and B.A.; formal analysis, D.D.K. and D.V.; investigation, D.D.K. and D.V.; resources, M.I.M., D.D.K. and T.P.; writing—original draft preparation, M.I.M., I.B., D.V., D.D.K., M.K., B.A. and T.P.; writing—review and editing, I.B. and D.V.; visualisation, M.K. and B.A.; supervision, I.B and T.P. All authors have read and agreed to the published version of the manuscript.

Funding

This research was funded by the Provincial Secretariat for Higher Education and Scientific Research, Autonomous Province of Vojvodina, Republic of Serbia, grant number 142-451-3480/2023-01 (Project Title: Investigation of dentists’ exposure to microbiological agents from respiratory aerosols during dental interventions).

Institutional Review Board Statement

Not applicable.

Informed Consent Statement

Not applicable.

Data Availability Statement

The original contributions presented in the study are included in the article material; further inquiries can be directed to the corresponding author.

Conflicts of Interest

The authors declare no conflicts of interest.

References

  1. Matys, J.; Grzech-Leśniak, K. Dental Aerosol as a Hazard Risk for Dental Workers. Materials 2020, 13, 5109. [Google Scholar] [CrossRef] [PubMed]
  2. MDHS14/4; General Methods for Sampling and Gravimetric Analysis of Respirable, Thoracic and Inhalable Aerosols. Health and Safety Executive: Aberdeen, UK, 2014.
  3. Licina, D.; Boor, B.E. Size-resolved inhalation intake fractions for particles released from human activities in residential indoor environments. Build. Environ. 2023, 244, 110763. [Google Scholar] [CrossRef]
  4. Matys, J.; Gedrange, T.; Dominiak, M.; Grzech-Leśniak, K. Quantitative Evaluation of Aerosols Produced in the Dental Office during Caries Treatment: A Randomized Clinical Trial. J. Clin. Med. 2023, 12, 4597. [Google Scholar] [CrossRef] [PubMed]
  5. Limsiriwong, K.; Winijkul, E. Exploring Personal Exposure to Airborne Microplastics across Various Work Environments in Pathum Thani Province, Thailand. Int. J. Environ. Res. Public Health 2023, 20, 7162. [Google Scholar] [CrossRef] [PubMed]
  6. Sidibe, A.; Sakamoto, Y.; Murano, K.; Koita, O.A.; Traore, I.; Dansoko, Y.; Kajii, Y. Personal Exposure to Fine Particles (PM2.5) in Northwest Africa: Case of the Urban City of Bamako in Mali. Int. J. Environ. Res. Public Health 2022, 19, 611. [Google Scholar] [CrossRef] [PubMed]
  7. Ulusoy, U. A Review of Particle Shape Effects on Material Properties for Various Engineering Applications: From Macro to Nanoscale. Minerals 2023, 13, 91. [Google Scholar] [CrossRef]
  8. Avino, P.; Protano, C.; Vitali, M.; Manigrasso, M. Benchmark study on fine-mode aerosol in a big urban area and relevant doses deposited in the human respiratory tract. Environ. Pollut. 2016, 216, 530–537. [Google Scholar] [CrossRef]
  9. Lin, H.; Li, X.; Qin, X.; Cao, Y.; Ruan, Y.; Leunga, M.K.H.; Leung, K.M.Y.; Lam, P.K.S.; He, Y. Particle size-dependent and route-specific exposure to liquid crystal monomers in indoor air: Implications for human health risk estimations. Sci. Total Environ. 2024, 908, 168328. [Google Scholar] [CrossRef] [PubMed]
  10. Deng, Q.; Deng, L.; Miao, Y.; Guo, X.; Li, Y. Particle deposition in the human lung: Health implications of particulate matter from different sources. Environ. Res. 2019, 169, 237–245. [Google Scholar] [CrossRef] [PubMed]
  11. Vincent, J.H. Aerosol Sampling: Science, Standards, Instrumentation and Applications; Wiley & Sons: Chichester, UK, 2007; ISBN 978-0-470-02725-7. [Google Scholar]
  12. Alenezi, A.; Qureshi, H.; Ahmed, O.G.; Ramanathan, M. Air Quality, Allergic Rhinitis, and Asthma. Otolaryngol. Clin. N. Am. 2024, 57, 2. [Google Scholar] [CrossRef] [PubMed]
  13. Gehr, P.; Mühlfeld, C.; Rothen-Rutishauser, B.; Blank, F. Particle-Lung Interactions, 2nd ed.; CRC Press: Boca Raton, FL, USA; Taylor & Francis Group: Abingdon, UK, 2009; ISBN 978-1-4200-7256-3. [Google Scholar]
  14. Phalen, R.F.; Phalen, R.N. Introduction to Air Pollution Science, a Public Health Perspective; Jones & Bartlett Publishers: Burlington, MA, USA, 2012; ISBN 978-0-7637-8044-9. [Google Scholar]
  15. Martins, V.; Cruz Minguillón, M.; Moreno, T.; Querol, X.; Miguel, E.; Capdevil, M.; Centelles, S.; Lazaridis, M. Deposition of aerosol particles from a subway microenvironment in the human respiratory tract. J. Aerosol Sci. 2015, 90, 103–113. [Google Scholar] [CrossRef]
  16. Zemouri, C.; de Soet, H.; Crielaard, W.; Laheij, A. A scoping review on bio-aerosols in healthcare and the dental environment. PLoS ONE 2017, 12, e0178007. [Google Scholar] [CrossRef] [PubMed]
  17. Helmis, C.G.; Tzoutzas, J.; Flocas, H.A.; Halios, C.H.; Stathopoulou, O.I.; Assimakopoulos, V.D.; Panis, V.; Apostolatou, M.; Sgouros, G.; Adam, E. Indoor air quality in a dentistry clinic. Sci. Total Environ. Times 2007, 377, 349–365. [Google Scholar] [CrossRef] [PubMed]
  18. Zhou, M.; Ma, J.; Wang, D.; Chen, W. Epidemiological Study of Respiratory Diseases among Dust Exposed Workers. In Encyclopedia of Respiratory Medicine, 2nd ed.; Elsevier: Amsterdam, The Netherlands, 2022; pp. 438–449. ISBN 9780081027240. [Google Scholar]
  19. Rosa, V.; Agarwalla, S.V.; Tan, B.L.; Choo, S.Y.; Sim, Y.F.; Boey, F.Y.C.; Anantharaman, S.; Duggal, M.S.; Tan, K.S. Pandemic Preparedness and Response: A Foldable Tent to Safely Remove Contaminated Dental Aerosols—Clinical Study and Patient Experience. Appl. Sci. 2022, 12, 7409. [Google Scholar] [CrossRef]
  20. Tang, F.; Wen, X.; Zhang, X.; Qi, S.; Tang, X.; Huang, J.; Zhu, C.; Shang, G.; Xu, Y.; Cai, J.; et al. Ultrafine particles exposure is associated with specific operative procedures in a multi-chair dental clinic. Heliyon 2022, 8, e11127. [Google Scholar] [CrossRef] [PubMed]
  21. Polednik, B. Aerosol and bioaerosol particles in a dental office. Environ. Res. 2014, 134, 405–409. [Google Scholar] [CrossRef] [PubMed]
  22. Godwin, C.C.; Batterman, S.A.; Sahni, S.P.; Peng, C.Y. Indoor environment quality in dental clinics: Potential concerns from particulate matter Comparative Study. Am. J. Dent. 2003, 16, 260–266. [Google Scholar] [PubMed]
  23. Schraufnage, D.E. The health effects of ultrafine particles. Exp. Mol. Med. 2020, 52, 311–317. [Google Scholar] [CrossRef] [PubMed]
  24. Lang, A.; Ovsenik, M.; Verdenik, I.; Remskar, M.; Oblak, Č. Nanoparticle concentrations and composition in a dental office and dental laboratory: A pilot study on the influence of working procedures. J. Occup. Environ. Hyg. 2018, 15, 441–447. [Google Scholar] [CrossRef] [PubMed]
  25. Polednik, B. Exposure of staff to aerosols and bioaerosols in a dental office. Build. Environ. 2021, 187, 107388. [Google Scholar] [CrossRef]
  26. Syed, M.; Chopra, R.; Sachdev, V. Allergic reactions to dental materials—A systematic review. J. Clin. Diagn. Res. 2015, 9, ZE04–ZE09. [Google Scholar] [CrossRef]
  27. Valente, L.L.; Peralta, S.L.; Ogliari, F.A.; Cavalcante, L.M.; Moraes, R.R. Comparative evaluation of dental resin composites based on micron- and submicron-sized monomodal glass filler particles. Dent. Mater. J. 2013, 29, 11. [Google Scholar] [CrossRef] [PubMed]
  28. Ishikawa, S.; Ishikawa, H.; Shindo, T.; Yoshida, T.; Shimoyama, Y.; Satomi, T.; Fujii, S.; Hamamoto, Y.; Iino, M.; Fukao, A. Effects of occupational environmental controls and work management on chromosomal damage in dental technicians in Japan. Int. J. Hyg. Environ. Health 2013, 216, 100–107. [Google Scholar] [CrossRef] [PubMed]
  29. Kim, T.S.; Kim, H.A.; Heo, Y.; Park, Y.; Park, C.Y.; Roh, Y.M. Level of Silica in the Respirable Dust Inhaled by Dental Technicians with Demonstration of Respirable Symptoms. Ind. Health 2002, 40, 260–265. [Google Scholar] [CrossRef] [PubMed]
  30. Al Jabbari, Y.S.; Koutsoukis, T.; Barmpagadaki, X.; Zinelis, S. Metallurgical and interfacial characterization of PFM Co–Cr dental alloys fabricated via casting, milling or selective laser melting. Dent. Mater. 2014, 30, 4. [Google Scholar] [CrossRef] [PubMed]
  31. Checchi, L.; Montevecchi, M.; Moreschi, A.; Graziosi, F.; Taddei, P.; Violante, F.S. Efficacy of three face masks in preventing inhalation of airborne contaminants in dental practice. J. Am. Dent. Assoc. 2005, 136, 877–882. [Google Scholar] [CrossRef]
  32. Manarte-Monteiro, P.; Carvalho, A.; Pina, C.; Oliveira, H.; Manso, M.C. Air quality assessment during dental practice: Aerosols bacterial counts in an universitary clinic. Rev. Port. Estomatol. Med. Dent. Cir. Maxilofac. 2013, 54, 2–7. [Google Scholar] [CrossRef]
  33. Shekunov, B.Y.; Chattopadhyay, P.; Tong, H.H.; Chow, A.H. Particle size analysis in pharmaceutics: Principles, methods and applications. Pharm. Res. 2007, 24, 203–227. [Google Scholar] [CrossRef] [PubMed]
  34. Kunda, N.K.; Somavarapu, S.; Gordon, S.B.; Hutcheon, G.A.; Saleem, I.Y. Nanocarriers targeting dendritic cells for pulmonary vaccine delivery. Pharm. Res. 2013, 30, 325–341. [Google Scholar] [CrossRef] [PubMed]
  35. Demoly, P.; Hagedoorn, P.; de Boer, A.H.; Frijlink, H.W. The clinical relevance of dry powder inhaler performance for drug delivery. Respir. Med. 2014, 108, 1195–1203. [Google Scholar] [CrossRef] [PubMed]
  36. Dominique, N.P.; Nitesh, K.K.; Pavan, M. Challenges associated with the pulmonary delivery of therapeutic dry powders for preclinical testing KONA. Powder Part. J. 2019, 36, 129–144. [Google Scholar] [CrossRef]
  37. Feng, Y.; Kleinstreuer, C. Analysis of non-spherical particle transport in complex internal shear flows. Phys. Fluids 2013, 25, 9. [Google Scholar] [CrossRef]
  38. Liu, M.H.; Chen, C.T.; Chuang, L.C.; Lin, W.M.; Wan, G.H. Removal efficiency of central vacuum system and protective masks to suspended particles from dental treatment. PLoS ONE 2019, 14, e0225644. [Google Scholar] [CrossRef] [PubMed]
  39. Nulty, A.; Lefkaditis, C.; Zachrisson, P.; Van Tonder, Q.; Yar, R. A clinical study measuring dental aerosols with and without a high-volume extraction device. Br. Dent. J. 2020, 20, 1–8. [Google Scholar] [CrossRef]
  40. World Health Organization. WHO/ILO Joint Estimates of the Work-Related Burden of Disease and Injury, 2000–2016: Global Monitoring Report 2021; World Health Organization: Geneva, Switzerland, 2021. [Google Scholar]
  41. Chen, J.; Shi, X.; Liu, S.; Jiang, Z.; Wang, M. Study on the factors influencing the performance of dust collection hood of angle grinder based on CFD-DEM simulation. Adv. Powder Technol. 2024, 35, 104346. [Google Scholar] [CrossRef]
  42. Balanta-Melo, J.; Guti’errez, A.; Sinisterra, G.; Díaz-Posso, M.d.M.; Gallego, D.; Villavicencio, J.; Contreras, A. Rubber dam isolation and high-volume suction reduce ultrafine dental aerosol particles: An experiment in a simulated patient. Appl. Sci. 2020, 10, 6345. [Google Scholar] [CrossRef]
  43. Jiang, Y.; Wingert, N.; Arif, A.; Garcia-Käufer, M.; Schulz, S.D.; Hellwig, E.; Gminski, R.; Polydorou, O. Cytotoxic and inflammatory response of human lung epithelial cells A549 to particles released from dental restorative materials during dry and wet grinding. Dent. Mater. J. 2022, 38, 1886–1899. [Google Scholar] [CrossRef] [PubMed]
  44. Wang, J.; Tian, Y.; Hu, X.; Han, J.; Liu, B. Integrated assessment and optimization of dual environment and production drivers in grinding. Energy 2023, 272, 127046. [Google Scholar] [CrossRef]
  45. Ilić, M.; Budak, I.; Vučinić Vasića, M.; Nagode, A.; Kozmidis-Luburić, U.; Hodolič, J.; Puškar, T. Size and shape particle analysis by applying image analysis and laser diffraction—Inhalable dust in a dental laboratory. Measurement 2015, 66, 109–117. [Google Scholar] [CrossRef]
  46. DeCost, B.L.; Holm, E.A. A large dataset of synthetic SEM images of powder materials and their ground truth 3D structures. Data Brief 2016, 9, 727–731. [Google Scholar] [CrossRef] [PubMed]
  47. Kaltofen, M.; Gundersen, A. Radioactively-hot particles detected in dusts and soils from Northern Japan by combination of gamma spectrometry, autoradiography, and SEM/EDS analysis and implications in radiation risk assessment. Sci. Total Environ. 2017, 607–608, 1065–1072. [Google Scholar] [CrossRef] [PubMed]
  48. Elmes, M.; Gasparon, M. Sampling and single particle analysis for the chemical characterisation of fine atmospheric particulates: A review. J. Environ. Manag. 2017, 202, 137–150. [Google Scholar] [CrossRef] [PubMed]
  49. Bouwman, A.M.; Bosmaa, J.C.; Vonk, P.; Wesselingh, J.A.; Frijlink, H.W. Which shape factor(s) best describe granules? Powder Technol. 2004, 146, 66–72. [Google Scholar] [CrossRef]
  50. Panagiotakisa, C.; Argyros, A. Parameter-free modelling of 2D shapes with ellipses. Pattern Recognit. 2016, 53, 259–275. [Google Scholar] [CrossRef]
  51. Bagheri, G.H.; Bonadonna, C.; Manzella, I.; Vonlanthen, P. On the characterization of size and shape of irregular particles. Powder Technol. 2015, 270, 141–153. [Google Scholar] [CrossRef]
  52. Petrak, D.; Dietrich, S.; Eckardt, G.; Köhler, M. Two-dimensional particle shape analysis from chord measurements to increase accuracy of particle shape determination. Powder Technol. 2015, 284, 25–31. [Google Scholar] [CrossRef]
  53. de Oliveira Morais, P.A.; de Souza, D.M.; de Melo Carvalho, M.T.; Madari, B.E.; de Oliveira, A.E. Predicting soil texture using image analysis. Microchem. J. 2019, 146, 455–463. [Google Scholar] [CrossRef]
  54. Li, B.; Yu, W.; Huang, H.M.; Lim, E.; Boiarkina, L.; Young, B. Struvite image analysis and its application for product purity prediction. J. Environ. Chem. Eng. 2019, 7, 103349. [Google Scholar] [CrossRef]
  55. Williams, K.C.; Chen, W.; Weeger, S.; Donohue, T.J. Particle shape characterisation and its application to discrete element modelling. Particuology 2014, 12, 80–89. [Google Scholar] [CrossRef]
  56. ISO 9276-6:2008; Representation of Results of Particle Size Analysis, Part 6: Descriptive and Quantitative Representation of Particle Shape and Morphology. International Organization for Standardization: Geneva, Switzerland, 2008.
  57. ASTM F1877-16; Standard Practice for Characterization of Particles. ASTM International: West Conshohocken, PA, USA, 2010.
  58. ISO 25178-2:2021; Geometrical Product Specifications (GPS) Surface Texture: Areal Part 2: Terms, Definitions and Surface Texture Parameters. International Organization for Standardization: Geneva, Switzerland, 2021.
  59. Yunker, M.B.; Belicka, L.L.; Harvey, H.R.; Macdonald, R.W. Tracing the inputs and fate of marine and terrigenous organic matter in Arctic Ocean sediments: A multivariate analysis of lipid biomarkers. Deep-Sea Res. II Top. Stud. Oceanogr. 2005, 52, 3478–3508. [Google Scholar] [CrossRef]
  60. Shawky, E.; Selim, D. Evaluation of the effect of extraction solvent and organ selection on the chemical profile of Astragalus spinosus using HPTLC-multivariate image analysis. J. Chromatogr. B 2017, 1061–1062, 134–138. [Google Scholar] [CrossRef] [PubMed]
  61. Makvandi, S.; Beaudoin, G.; McClenaghan, M.B.; Quirt, D.; Ledru, P. PCA of Fe-oxides MLA data as an advanced tool in provenance discrimination and indicator mineral exploration: Case study from bedrock and till from the Kiggavik U deposits area (Nunavut, Canada). J. Geochem. Explor. 2019, 197, 199–211. [Google Scholar] [CrossRef]
Applsci 14 05169 g001
Figure 1. Methodology of research.
Figure 1. Methodology of research.
Applsci 14 05169 g001
Applsci 14 05169 g002
Figure 2. Dental technician with personal sampler.
Figure 2. Dental technician with personal sampler.
Applsci 14 05169 g002
Applsci 14 05169 g003
Figure 3. SEM micrographs.
Figure 3. SEM micrographs.
Applsci 14 05169 g003
Applsci 14 05169 g004
Figure 4. Parameter adjustments in the ImageJ software: (a) original image; (b) after applying the Gaussian filter; (c) threshold adjustment.
Figure 4. Parameter adjustments in the ImageJ software: (a) original image; (b) after applying the Gaussian filter; (c) threshold adjustment.
Applsci 14 05169 g004
Applsci 14 05169 g005
Figure 5. Measurement setup and image analysis results in ImageJ software.
Figure 5. Measurement setup and image analysis results in ImageJ software.
Applsci 14 05169 g005
Applsci 14 05169 g006
Figure 6. Parameter adjustment in the JMicroVision software: (a) threshold adjustment; (b) particle selection.
Figure 6. Parameter adjustment in the JMicroVision software: (a) threshold adjustment; (b) particle selection.
Applsci 14 05169 g006
Applsci 14 05169 g007
Figure 7. Results in JMicroVision software.
Figure 7. Results in JMicroVision software.
Applsci 14 05169 g007
Applsci 14 05169 g008
Figure 8. Probability plot testing for normal distribution.
Figure 8. Probability plot testing for normal distribution.
Applsci 14 05169 g008
Applsci 14 05169 g009
Figure 9. Scree method of principal component analysis of geometric parameters.
Figure 9. Scree method of principal component analysis of geometric parameters.
Applsci 14 05169 g009
Applsci 14 05169 g010
Figure 10. Biplot of principal component analysis of geometric parameters of the sample.
Figure 10. Biplot of principal component analysis of geometric parameters of the sample.
Applsci 14 05169 g010
Applsci 14 05169 g011
Figure 11. Dendrogram of the investigated geometric parameters.
Figure 11. Dendrogram of the investigated geometric parameters.
Applsci 14 05169 g011
Table 1. Definitions of the geometry parameters [56,57,58] used to characterise the particles in this study.
Table 1. Definitions of the geometry parameters [56,57,58] used to characterise the particles in this study.
ParametersAbbreviationDefinition
SizeAreaAArea of a particle using the two-dimensional (2D) projection.
PerimeterPContour length (P), i.e., the total length of the boundary of the 2D projection of the particle.
LengthLCalliper length along the orientation axis of the object parallel to the longest axis.
WidthWCalliper length along the orientation axis + 90° of the object parallel to the shortest axis.
MeanFeret’s diameter D ¯ F The Feret’s diameter (dF) is the longest distance between any two points along the selection boundary, also known as maximum calliper. The minimal Feret’s diameter (Min dF) is the minimal Feret’s diameter calculated after considerations of all possible orientations (0°…180°). We use mean values.
Equivalent Circle DiameterDCEDiameter of a circle with area equivalent to area of a particle using the 2D projection.
Shape
Morphological roughness		SoliditySLDRatio of the area of the object to the convex area.
CompactnessCPRatio of the area of the object to the area of a circle with the same perimeter.
Shape
Textural roughness		ConvexityCVXRatio of the convex perimeter to the perimeter of the object.
RectangularityRTRatio of the area of a rectangle (formed with length and width as sides) to the area of the object.
Shape
Form		ElongationElRatio of the length to the width.
Aspect RatioARThis is the length of the major axis divided by the length of the minor axis of the fitted ellipse.
RoundnessRDThe roundness of the particle, defined as: 4 × area/π × (major axis)2.
Shape
Form and roughness		CircularityCirc4π (area/perimeter2).
Table 2. Pearson correlation coefficient of geometric parameters.
Table 2. Pearson correlation coefficient of geometric parameters.
Particle Size ParametersParticle Shape Parameters
AP D ¯ F LWDCEElRTCVXCircARRDSLDCP
A1.000
P0.6931.000
D ¯ F 0.7550.9821.000
L0.7770.9680.9871.000
W0.6650.9430.9550.9081.000
DCE0.7670.9610.9890.9720.9681.000
El−0.081−0.244−0.231−0.311−0.056−0.1811.000
RT0.1180.5310.4930.4860.5080.431−0.2441.000
CVX−0.143−0.589−0.489−0.471−0.516−0.452−0.220−0.7321.000
Circ−0.138−0.567−0.506−0.509−0.472−0.4530.501−0.7890.8511.000
AR0.1030.3380.3190.3990.1330.253−0.8500.322−0.283−0.5761.000
RD−0.101−0.328−0.309−0.376−0.149−0.2550.932−0.3570.3260.617−0.9051.000
SLD−0.036−0.329−0.269−0.284−0.220−0.1910.402−0.7030.6080.793−0.4700.5141.000
CP−0.140−0.552−0.512−0.511−0.501−0.4760.444−0.8230.8270.940−0.5130.5540.6891.000
Table 3. Latent root criterion of principal component analysis of geometric parameter.
Table 3. Latent root criterion of principal component analysis of geometric parameter.
Eigenvalue8.9382.7331.5220.3140.1870.0970.0730.0650.0370.0230.0050.0040.0030.002
Proportion0.6380.1950.109 0.0220.0130.0070.0050.0050.0030.002 0.0000.0000.000 0.000
Cumulative0.6380.8330.9420.9650.9780.9850.9900.9950.9970.9990.999 1.0001.0001.000
Table 4. PCA of geometric parameters.
Table 4. PCA of geometric parameters.
Geometric ParametersComponents
PC1PC2PC3
A−0.2950.288−0.216
P−0.3180.151−0.111
D ¯ F 0.308−0.1830.173
L0.318−0.1150.177
W−0.2850.302−0.077
DCE−0.2940.230−0.222
El0.1540.4680.297
RT−0.2820.0160.317
CVX0.2640.051−0.381
Circ0.2920.174−0.273
AR0.1480.4610.320
RD0.1900.4440.231
SLD0.1980.254−0.466
CP0.3040.088−0.248
Table 5. PCA of geometric parameters after rotation.
Table 5. PCA of geometric parameters after rotation.
Geometric ParametersFactors
F1F2F3
A−0.001−0.429−0.042
P−0.019−0.3620.071
D ¯ F 0.0200.397−0.011
L0.0820.372−0.030
W0.132−0.4000.039
DCE−0.003−0.431−0.047
El0.5750.0040.014
RT0.091−0.0820.407
CVX−0.0730.007−0.460
Circ0.0890.012−0.426
AR0.5790.0130.037
RD0.5330.015−0.051
SLD0.044−0.174−0.520
CP0.0350.076−0.393
Disclaimer/Publisher’s Note: The statements, opinions and data contained in all publications are solely those of the individual author(s) and contributor(s) and not of MDPI and/or the editor(s). MDPI and/or the editor(s) disclaim responsibility for any injury to people or property resulting from any ideas, methods, instructions or products referred to in the content.

Share and Cite

MDPI and ACS Style

Ilic Micunovic, M.; Budak, I.; Vukelic, D.; Djurovic Koprivica, D.; Kuzmanovic, M.; Agarski, B.; Puskar, T. Investigation of the Geometric Characteristics of Inhalable Particles Emitted from the Process of Grinding Dental Restorations. Appl. Sci. 2024, 14, 5169. https://doi.org/10.3390/app14125169

AMA Style

Ilic Micunovic M, Budak I, Vukelic D, Djurovic Koprivica D, Kuzmanovic M, Agarski B, Puskar T. Investigation of the Geometric Characteristics of Inhalable Particles Emitted from the Process of Grinding Dental Restorations. Applied Sciences. 2024; 14(12):5169. https://doi.org/10.3390/app14125169

Chicago/Turabian Style

Ilic Micunovic, Milana, Igor Budak, Djordje Vukelic, Daniela Djurovic Koprivica, Milos Kuzmanovic, Boris Agarski, and Tatjana Puskar. 2024. "Investigation of the Geometric Characteristics of Inhalable Particles Emitted from the Process of Grinding Dental Restorations" Applied Sciences 14, no. 12: 5169. https://doi.org/10.3390/app14125169

Note that from the first issue of 2016, this journal uses article numbers instead of page numbers. See further details here.

Article Metrics

Back to TopTop