Journal of Spectroscopy

Journal of Spectroscopy / 2018 / Article

Research Article | Open Access

Volume 2018 |Article ID 5837214 |

Surendra P. Verma, Sanjeet K. Verma, M. Abdelaly Rivera-Gómez, Darío Torres-Sánchez, Lorena Díaz-González, Alejandra Amezcua-Valdez, Beatriz Adriana Rivera-Escoto, Mauricio Rosales-Rivera, John S. Armstrong-Altrin, Héctor López-Loera, Fernando Velasco-Tapia, Kailasa Pandarinath, "Statistically Coherent Calibration of X-Ray Fluorescence Spectrometry for Major Elements in Rocks and Minerals", Journal of Spectroscopy, vol. 2018, Article ID 5837214, 13 pages, 2018.

Statistically Coherent Calibration of X-Ray Fluorescence Spectrometry for Major Elements in Rocks and Minerals

Academic Editor: Rafal Sitko
Received08 Aug 2018
Revised08 Oct 2018
Accepted19 Oct 2018
Published11 Dec 2018


We applied both the ordinary linear regression (OLR) and the new uncertainty weighted linear regression (UWLR) models for the calibration and comparison of a XRF machine through 59 geochemical reference materials (GRMs) and a procedure blank sample. The mean concentration and uncertainty data for the GRMs used for the calibrations (Supplementary Materials) (available here) filewere achieved from an up-to-date compilation of chemical data and their processing from well-known discordancy and significance tests. The drift-corrected XRF intensity and its uncertainty were determined from mostly duplicate pressed powder pellets. The comparison of the OLR (linear correlation coefficient ∼0.9523–0.9964 and 0.9771–0.9999, respectively, for before and after matrix correction) and UWLR models (∼0.9772–0.9976 and 0.9970–0.9999, respectively) clearly showed that the latter with generally higher values of is preferable for routine calibrations of analytical procedures. Both calibrations were successfully applied to rock matrices, and the results were generally consistent with those obtained in other laboratories although the UWLR model showed mostly narrower confidence limits of the mean (slope and intercept) or lower uncertainties than the OLR. Similar sensitivity (∼2.69–46.17 kc·s−1·%−1 for the OLR and ∼2.78–59.69 kc·s−1·%−1 for the UWLR) also indicated that the UWLR could advantageously replace the OLR model. Another novel aspect is that the total uncertainty can be reported for individual chemical data. If the analytical instruments were routinely calibrated from the UWLR model, this action would make the science of geochemistry more quantitative than at present.

1. Introduction

All modern analytical instruments require some kind of calibration of the instrumental response (-variable) as a function of the concentration (-variable) [13]. This calibration is generally achieved through an ordinary least-squares linear regression (OLR) model. However, such a procedure is not strictly valid because all requirements for the statistical validity of the OLR model are not fulfilled. Usually, the assumptions “independent concentration variable is error-free or less than one-tenth of the error in the dependent response variable ” and “error in is homoscedastic” (i.e., equal errors for all values) are not satisfied and, therefore, more sophisticated and statistically coherent regression procedures, such as weighted least-squares linear regression (WLR) models, should be used [418].

X-ray fluorescence (XRF) spectrometry is among the most popular analytical techniques for the determination of all major and some trace elements in rocks [4, 1927]. Natural geochemical reference materials (GRMs) are commonly used for XRF calibrations and posterior characterization of those and other GRMs as well as of similar rock and mineral matrices [4, 19, 2830]. As for most other analytical instruments, XRF spectrometers are also calibrated under the statistically incoherent OLR model.

To apply the WLR and compare it with the OLR, both central tendency (e.g., mean) and dispersion (e.g., confidence limits of the mean) estimates on both -axis (concentration, generally expressed in the unit of % m/m, i.e., mass/mass unit expressed in percent) and -axis (response, in this case XRF intensity, generally reported in the unit of kc·s−1, i.e., kilo counts per second) variables are required. More precise (and accurate) estimates of the central tendency will also be useful for both types of regressions. Therefore, precise concentrations of GRMs with the respective lowest possible “confidence limits of the mean” (referred hereafter as the “uncertainty” of the measured variable) [2, 17, 18] are required to apply the regression procedure. Sometimes, we had to use also the term “error” (instead of the uncertainty) because the use of the error is widespread in the literature.

We report the following five aspects: (a) evaluation of 59 GRMs to achieve the least possible uncertainties in the mean concentrations of all major elements (SiO2 to P2O5); (b) the comparison of regression models (OLR and WLR) applied to net drift-corrected XRF intensities before the correction of matrix effects; (c) the second (or final) comparison of both models after achieving the matrix correction as well as for the estimation of sensitivities of the regression models; (d) application of the entire procedure to four GRMs treated as “unknown” samples and their comparison with the previous literature compilations; and (e) development of a computer program to achieve the abovementioned objectives. Thus, the regression equations (intercept and its uncertainty, slope and its uncertainty, and linear correlation coefficient values) for each constituent from SiO2 to P2O5 and their application to similarly complex rock matrices are presented in this work.

2. Evaluation of Major Element Data for GRMs

A total of 59 GRMs (listed in alphabetical order in Table S1; this and four other tables are provided in Supplementary Materials), along with a procedure blank, were used in this study. The procedure blank was a pellet prepared in duplicate with only pure N,N′-Ethylene bis(stearamide) beads without any sample (Section 3). The individual data reported in earlier compilations [3147] were first compiled in new databases.

The statistical parameters obtained in these early compilations could not be directly used for instrumental calibrations due to the following reasons: (i) the statistical methods used to achieve the statistical estimates were outdated (see [17, 18, 48, 49] for possible reasons), and the inferred statistical values were of low quality (high values of dispersion); (ii) there are still determinations reported during about 30 or more years (postcompilation years) that were not obviously available to those compilers; (iii) the precision of more recent determinations is likely to have improved due to the availability of online computers on most modern instruments; (iv) newer more reliable statistical techniques are now available for improving both precision and accuracy of the statistical inferences, e.g., the use of discordancy tests with the highest power and lowest swamping and masking effects [18, 48, 5052]; and (v) importantly, new computer programs have been developed by our group [5254], available at for download or online processing of data (after previous registration onto our server), which can be advantageously used for efficient processing of experimental databases.

The same kinds of objections are applicable even today for the originator’s websites, such as for Japanese GRMs or for United States GRMs. The statistical information at these websites is based on early compilations (around 30 or more years ago). Furthermore, we were unable to use the recent work [55] because this paper reported significantly larger uncertainty values as compared to those achievable from our new validated statistical procedure [5154]; besides, updated statistical information on the mean and its uncertainty was not available in [55] for many GRMs used in our work.

The initial databases were complemented by individual data from a large number of posterior publications (∼480; Table S1), whose complete listing is available at our server under the heading of “Quality Control.” These major element data were classified according to the analytical method groupings [56]. Data from each method group were considered as a univariate statistical sample. Appropriate discordancy and significance tests were applied from thoroughly automatized software UDASys2 [52] and UDASys3 (unpublished), which, in their “recommended procedure,” apply the most powerful five (two new and three conventional) recursive tests with prior application of respective single-outlier tests having nil swamping and low masking effects [48, 5760]. Although the application of discordancy tests is identical for both UDASys2 and UDASys3, the difference lies in that the latter applies the significance (ANOVA, F and t) tests in order to provide the final results automatically.

The resulting statistical information after the application of well-known discordancy tests at the strict 99% confidence level (mean and uncertainty values rounded according to the flexible rules [18]) is listed in Table S2. These GRM compositional data showed by far the lowest 99% uncertainty (Table S2), much lower than any existing compilation [3147, 55]. We may also stress once again that this was achieved through an objective combination of discordancy tests having the highest performance and lowest swamping and masking effects [17, 18, 48, 53], i.e., from the methodology having the lowest type I and type II errors and the highest power.

Therefore, the population mean of these GRMs is now known within the narrowest possible 99% confidence limits of the mean to best represent the concentration (x) axis in the instrumental calibrations as suggested [2, 5, 7, 10, 17, 18, 53]. These data (in units of % m/m; Table S2) will also be useful for those who wish to achieve instrumental calibrations or simply use them for quality control of their results for rock and mineral matrices.

3. XRF Instrumentation and Intensity Measurements

A wavelength dispersive X-ray fluorescence (WDXRF) spectrometer Rigaku ZSX Primus II model (rhodium X-ray tube; 4 kW maximum power) was used for this work. We made the effort to best represent the response () axis (-ray intensity in the units of kilo counts per second, kc·s−1) for the calibrations. For each GRM, duplicate (41 samples) or even triplicate (8 samples) pressed powder pellets were prepared. First, an appropriate amount of each GRM was dried overnight in an oven at about 105°C. For each pellet, accurately weighed 3.5 g of moisture-free GRM was thoroughly mixed with accurately weighed 3.0 g pure N,N′-ethylene bis(stearamide) beads, <840 μm as wax (Sigma-Aldrich), and stored in a desiccator. Pressed powder pellets were prepared at 20 tons·inch−2 pressure (about 310 MPa). However, for 10 GRMs, sufficient material was not available; therefore, only a single pellet could be prepared but the measured intensity uncertainty () at the 99% confidence level was increased by a factor of 2 to take into account the sample preparation variance. Similarly, accurately weighed 6.5 g of pure N,N′-ethylene bis(stearamide) beads, <840 μm as wax, was pressed to prepare a procedure blank sample. This was done in duplicate.

For the intensity measurements, the optimum instrumental conditions were first established through preliminary experiments prior to the routine measurements (Table S2). Each pellet was run at least 8 to 10 times in a random sequence, along with two drift monitors prepared from two volcanic rocks (basalt and rhyolite) from the San Luis Potosí Volcanic Field, San Luis Potosí (central Mexico).

The peak and background measuring conditions and time periods are also listed in Table S3. Appropriate mean drift corrections from two monitors were applied to all intensity measurements. Both monitors were run randomly 8 to 10 times each day. First, the expected monitor intensity was established as an average value of the first two days when the intensities were fairly stable and reproducible. Then, the average drift correction factors were calculated for each chemical element from the two monitors run in the XRF instrument periodically before and after a set of GRMs used for the calibration. These correction factors were then applied to the bracketed GRMs for the entire period of calibration, including the first two days and analysis of “unknown” samples.

Now, although the X-ray counts may obey a Poisson distribution, we are dealing with average values of count rates, which are likely to follow a normal distribution because of the central limit theorem. A normal distribution of measured intensities was also assured for each pellet from the application of discordancy tests as explained above for GRM concentrations. The intensity results for all pellets from a given GRM were then combined, the tests applied again to the combined data, and new mean and 99% uncertainty values were calculated for X-ray intensity of each GRM. This was done to take into account the variance of the sample preparation method, which was significantly higher than the instrumental variance of intensity measurements for individual pellets. The drift-corrected intensity values and their 99% uncertainties (kc·s−1) for all GRMs, along with the concentration data and their 99% uncertainties (% m/m), are listed in Table S2.

4. Regression Models

Two different regression models (OLR and UWLR) were used and compared in this work. The OLR model most frequently used for instrumental calibrations (-axis concentration and -axis response; GRM concentration and X-ray intensity, respectively, in XRF spectrometry) requires the following assumptions to be fulfilled [4, 7, 10, 1218]: (i) all errors are in the -axis; (ii) -axis is either error-free or has at most 10% error of the -axis errors; (iii) errors in both axes are normally distributed; and (iv) errors in the -axis are homoscedastic. Some or all of these assumptions are violated in most instrumental calibrations through the OLR model.

Thus, from the literature on the GRMs, it has been demonstrated that the concentration axis is not error-free (see non-zero uncertainties for all GRM concentrations in Table S2) [3147, 5153]. One can also clearly see that the errors in the intensity axis are not homoscedastic (see unequal, i.e., heteroscedastic uncertainties for any element in different GRMs in Table S2). For a heteroscedastic linear regression system, even if each error or noise term is still Gaussian, the OLR model is no longer the maximum likelihood estimate and consequently, it is no longer efficient [10]. The main advantage that the WLR has over the OLR is the ability to handle regression situations in which the data points are of varying quality as is the case with most instruments including the XRF spectrometers.

However, the major disadvantage of the WLR is that the approach is based on the assumption that the weights are known exactly. They can be estimated using several different equations or algorithms, but when the weights are produced from small numbers of replicated observations, the regression parameters can be unpredictably affected [10]. In the example of the XRF calibration that we are presenting, the numbers of observations were relatively large for both the and axes (concentration and X-ray intensity parameters). Besides, instead of the sample variance, we used the uncertainty values (that take into account the number of observations in the formula for uncertainty or confidence limits of the mean calculations) [2, 18] for estimating the weight factors. The problem of the sensitivity to outliers in the regression equations [10] was also appropriately handled by discordancy tests programmed in the UDASys and BiDASys software [53, 54, 61].

Therefore, although frequently used, the OLR model is not statistically correct or coherent. The statistically coherent WLR, especially the uncertainty-based WLR (UWLR [17]) model, should be used. The confidence level, such as 95% or 99% (significance level of 5% or 1%, respectively, or of 0.05 and 0.01, respectively), can be explicitly expressed in the confidence limits of the mean or uncertainty used in the UWLR model as well as to estimate the weight factors [17]. We will deal with the 99% uncertainty to have the type I error small (about 1%). Unfortunately, software of most analytical instruments, including XRF spectrometers, allows only the OLR calibration. Therefore, any sophisticated regression model, such as the UWLR, will have to be applied outside the instrumental software. Thus, the probability concept (99% confidence level) can be explicitly used in the UWLR model for weight factors based on the inverse of the squared 99% uncertainty of the mean.

We now present a synthesis of the regression equations for instrumental calibrations [2, 10, 17, 18, 61].

4.1. Ordinary Least-Squares Linear Regression (OLR) Model

Let us assume that we have a series of reference materials or standard calibrators having individual mean concentrations with respective uncertainties where varies from to . In order to calibrate an instrument, each of these calibrators were run several times, obtaining individual mean responses with respective uncertainties where varies from to . Thus, we have bivariate concentration-response data pairs or calibrators () with the respective uncertainties ().

We can apply the OLR model to these data for obtaining a calibration equation. The OLR fits a least-squares linear equation to the pairs () but does not take into account the respective uncertainties ().

The general regression equation for the OLR is as follows (the subscript is for the OLR model):where is the slope, is the resulting uncertainty in the slope, is the intercept, is the resulting uncertainty in the intercept, is the independent variable, is the dependent variable from the OLR model, and is the resulting uncertainty in . The following equations allow the calculations of these parameters:where and are, respectively, the mean values of the and variables:where is the value of for in equation (1) and is the Student’s t test value for degrees of freedom, and the superscript is the confidence level, generally 95% or 99%:

It is a general practice in most instrumental calibrations to ignore all uncertainties in equation (1) and use an OLR equation without any error (or uncertainty) as follows:

The resulting standard deviation values of repeat measurements of unknown samples are reported as the final errors. However, these are only partial errors because the errors in the calibration equation (1) are not taken into account. In this work, we will use equation (1) to report total errors (in fact, 99% uncertainties) for the OLR model.

4.2. Uncertainty Weighted Least-Squares Linear Regression (UWLR) Model

For the UWLR model, the pairs () of calibrators as well as the respective uncertainties () are taken into account in order to achieve the best least-squares linear fit.

The uncertainties in the x-axis are first propagated to the y-axis, combined with the , and the total uncertainty values on the y-axis are used for the weighting factors [2, 10, 17, 18, 61]:

The weights are calculated from as follows:where values have the following property:

Thus, the UWLR fits a linear equation to the pairs () with the respective weighting factors as follows (the subscript is for the UWLR model):

Note that this regression line will pass closer to the data with lesser uncertainty . The intercept and slope variables and their uncertainties are calculated from the following equations:where and are, respectively, the weighted mean values of the and variables:where is the value of for in equation (9):

The best regression equation for a calibration curve should have the following characteristics (without distinguishing the subscripts and ): (i) intercept small approaching to zero; (ii) slope large; and (iii) both and small. Further, the quality of the regression, whether a calibration curve or any other bivariate relationship, is also expressed as the linear regression coefficient (; and , respectively, for the OLR and UWLR), which is ideally +1.00000 for a calibration curve [5, 18, 61].

5. Application of Regression Models for XRF Calibration

5.1. Original Drift-Corrected Net Intensities and GRM Concentrations: The First Set of Two Regression Equations for Each Element

The evaluations for both regression types on the drift-corrected net intensity-concentration (Int-Conc) relationships (Table S2) for all major elements from SiO2 to P2O5 were performed (Table S4), for which the new online software BiDASys was used [61] at BiDASys allows the application of the conventional OLR as well as the newly proposed UWLR model [17] and provides the output of all regression parameters in an Excel® file. Contrary to the common practice, we will refrain from showing the numerous - (variable is drift-corrected net intensity “Int” and variable is the GRM concentration “Conc”) plots. This is because Table S4 statistically quantifies the visual interpretation of such diagrams. The quality parameters (standard errors and , uncertainty and , and linear correlation coefficient and its squared value parameters) are reported in Table S4. Because we are using these several different quality parameters, the concern against the use of solely parameter [62] is not important for comparison purposes.

We will explain the implications of the statistical results for the first element SiO2; the statistics for other elements (Table S4) can be similarly understood. The OLR regression equation from the first row of statistical information in Table S4 is as follows (after the element , subscript is for the OLR and is for provisional concentration; note many decimal places are used for the regression variables in such equations, because these values are not final results, and we should not introduce rounding errors during the calculation stage):

Similarly, the UWLR equation from the second row of statistical information in Table S4 is as follows:

The implications of these regression equations can be understood from the comparison of the uncertainties of the intercept and slope, which are lower for the UWLR (equation (14)) than for the OLR (equation (13)). This means that the uncertainty of the calculated concentration will be lower for the UWLR than for the OLR. Correspondingly, the value for the UWLR (0.99004, ; ) is much higher than that for the OLR (0.95229, ; ; Table S4). Similar trend in the (and ) values was obtained for all other elements except MnO (Table S4).

5.2. Matrix-Effect-Corrected Intensities and GRM Concentrations: The Second Set of Two Regression Equations for Each Element

Matrix correction is certainly required because the abovementioned least-squares linear regression fits are far from “perfect” ( ≠ +1.00000; in fact, < 1; ; = 0.95229–0.99638 for the OLR and = 0.97715–0.99760 for the UWLR; Table S4). There is a vast literature on the subject of matrix effects in XRF and their correction procedures [6375]. In this study, the Lachance-Traill algorithm [73] was used for the matrix effect correction [63, 71]. This was done outside the XRF instrument software. In a review of the existing algorithms, Rousseau [63] showed that the Lachance-Traill algorithm could be considered as one of the most appropriate procedures for the matrix effect correction because other algorithms have limited application range or lack of accuracy. Thus, for each element from SiO2 to P2O5, a system of overdetermined equations was solved and the resulting alpha coefficients were used to correct all intensities for matrix effects.

From the alpha coefficients, matrix-corrected intensities and improved concentration values for the GRMs and their uncertainties were calculated iteratively under the condition that the convergence parameter (absolute relative difference of the GRM calculated and input concentrations) for each compositional constituent (SiO2 to P2O5) be minimized.

New regression equations for achieving the corrected concentrations were established from the relationship of the calculated GRM concentrations (ConcCalc) and the original GRM concentrations (Conc) given in Table S2, for which the online BiDASys software [61] was used at These equations can be formulated from the regression coefficient values given in Table S4 (see ConcCalc-Conc rows corresponding to the OLR and UWLR). Again, we will highlight their significance for SiO2 only.

The OLR regression equation from the third row of statistical information in Table S4 is as follows:where the subscripts and stand for the OLR model and calculated concentration (ConcCalc), respectively.

Similarly, the UWLR equation from the fourth row of statistical information in Table S4 is as follows:where the subscripts and stand for the UWLR model and calculated concentration (ConcCalc), respectively.

Equations (15) and (16) show that the concentration values from the UWLR would be more reliable (lesser uncertainty values in both intercept and slope) than the OLR model. The value is higher for the UWLR (0.99704, ; ; Table S4) than the OLR (0.97710, ; ).

After the matrix correction, in fact most regression equations are better because all and values are higher for both OLR and UWLR than without the correction (Table S4; Figure 1 for only). For the OLR, the matrix correction increased the values () from 0.95229–0.99638 () to 0.97710–0.99992 (). Similarly, for the UWLR, this increase was from 0.97715–0.99760 () to 0.99704–0.99993 (). Thus, after matrix correction, all values increased for both OLR and UWLR. For the UWLR, the values approached the ideal value of +1.00000 (Figure 1). One has to keep in mind that when the values are closer to the maximum possible value of 1 (the “ideal” fit), the improvement expressed by the actual (absolute) value of will apparently be small. However, as long as the value increases for the UWLR as compared to the OLR (Figure 1; Table S4), we can objectively infer that the UWLR is a better regression model than the OLR.

Before the matrix correction, the intercepts of the Int-Conc regression lines were closer to zero for the UWLR (range ∼−0.013 to +0.011) than for the OLR (range ∼−2.098 to +11.47) model (Table S4; Figure 2). The same is true for the intercept values (ConcCalc-Conc relationship) after the matrix correction (∼−0.025 to +0.021 for the UWLR and ∼−0.110 to +1.87 for the OLR).

Finally, the uncertainties on both intercept and slope parameters were mostly lower for the UWLR than the OLR (Table S4). We highlight these differences (lower uncertainties for the UWLR) from dimensionless (free of the measurement units) parameters and defined as follows:

Plots of these two parameters are presented in Figure 3. If , the will be positive, otherwise it will be negative. The same is true for . For the comparison of two models OLR and UWLR before the matrix correction, the uncertainty for the UWLR were lower than the OLR for 7 elements (positive and ), whereas for after the matrix correction, it was so for 8 elements (out of 10; Figure 3). The exceptions were for 3 elements Mno, CaO, and P2O5 (negative and ) for the uncertainties before matrix correction and for 2 elements MnO and MgO for those after the matrix correction (Figure 3). Even for the exceptions of the elements MnO and MgO, the UWLR values should be usable (Table S4), i.e., it is not actually necessary to resort to the OLR model for these two exceptions (2 out of 10 cases). Thus, we can use the UWLR model for all purposes.

6. Sensitivities of Major Elements

We calculated the sensitivities as the slope of the Conc-IntCorr (GRM concentrations of Table S2 and matrix-corrected intensities of Table S5; see Supplementary Materials at from the regression curve (line) for all 10 elements and for both models (Table 1). Because the values are significantly high (all >0.961, ; Table 1) and the residuals are randomly distributed (graphs not shown), the straight line is the most likely, statistically valid fit for the concentration-matrix-corrected intensity data [5, 17, 18]. Therefore, the slope of the regression line represents an average sensitivity value for a given element under the chosen working conditions (Table S3).

ElementRegressionNumber of data pairs (calibrators)Regression equation parametersQuality of regression equation
Variables -ModelIntercept ()Slope ()











, intercept; , standard error; , uncertainty at 99%; , slope; , linear correlation coefficient; , squared linear correlation coefficient.

The intercept values were closer to zero (zero being the theoretically ideal intercept) for the UWLR regression (∼−0.113 to +0.104; Table 1) as compared to the OLR (∼−47.8 to +12.3; Table 1). The sensitivity values represented by the slopes of the regression lines (Table 1) were generally similar for both models (∼2.69–46.17 kc·s−1·%−1 for the OLR and ∼2.78–59.69 kc·s−1·%−1 for the UWLR). The sensitivity actually depends on the measuring conditions (Table S3), which were the same for both models.

For the matrix-corrected intensity-concentration (IntCorr-Conc) regressions, the parameters are listed in Table 2. All intercepts for the UWLR model, without exception, were closer to zero as compared to the OLR model. This confirms the superiority of the UWLR model.

ElementRegressionNumber of data pairs (calibrators)Regression equation parametersQuality of regression equation
Variables -ModelIntercept ()Slope ()











, intercept; , standard error; , uncertainty at 99%; , slope; , linear correlation coefficient; , squared linear correlation coefficient.

7. Application to Rock Matrices

The calibrations achieved in this work (Table S4) were applied to the analysis of four GRMs (attapulgite or Fuller’s earth clay ATT1; bentonite clay CSB1; granite GH; and tonalite TLM1) taken as “unknown” samples. These GRMs, having similarly complex matrices as the calibration samples, were not included in the calibrations because their mean values were available only from early description or compilations (for ATT1 and CSB1 [76]; for GH [77]; and for TLM1 [78]). We were unsuccessful in complementing these “old” concentration values with newer ones for these GRMs. Therefore, these GRMs were used as unknown samples. They were analysed in exactly the same manner as the calibration samples.

All calculations for the unknown samples were done outside the instrumental software. The drift-corrected net intensities and the corresponding uncertainties were processed from the first set of two regression equations (Int-Conc OLR and UWLR models; Table S4) to obtain provisional concentration and uncertainty values. The provisional concentrations were then used to obtain matrix corrections for each sample. The method was iteratively applied with the newer concentrations to obtain the final calculated concentration values (Table 3). These calculated concentration values were used to compute the final mean concentrations () and 99% uncertainties of the mean () for each sample from the second sets of regression equations (ConcCalc-Conc, OLR and UWLR models; Table S4). The loss on ignition (LOI) was required to optimise the final results.

RockMethod (ref.)SiO2TiO2Al2O3Fe2O3tMnOMgOCaONa2OK2OP2O5





The results are listed in Table 3 and compared with the literature compilations [7577]. On the other hand, because 99% uncertainties were not reported in the original compilations, they were computed for the comparison from the standard deviation, number of determinations, and appropriate two-sided t values at 99% confidence level [2, 18].

Firstly, although the mean concentration values determined by the OLR and UWLR models showed a general agreement, the 99% uncertainty values (; Table 3) were generally lower for the UWLR models, which clearly indicates that this model should be used routinely, instead of the conventional OLR model. Secondly, there is also a general agreement among all mean values, especially for granite GH and tonalite TLM1. The two clay samples (ATT1 and CSB1) showed some differences with the preliminary values obtained by the originators of these GRMs [75]. These values for comparison were obtained in only one laboratory. The errors (uncertainties) reported in the literature were underestimated, because they did not include those resulting from the calibrations. Furthermore, the accuracy data of the originator’s laboratory were not reported [75], such as the results for established GRMs and their comparison to other laboratories.

8. Computer Program XRFCalcUnknown

An online computer program JSpectrom_XRFCalcUnknown will be available at our server for use for unknown samples, which will guide other users to achieve the UWLR calibration outside of the instrumental software and its routine application to unknown samples. This program incorporates the iteration process to achieve reliable concentrations as demonstrated in this work. Example input data files and a ReadMe document are provided to facilitate the application of JSpectrom_XRFCalcUnknown. One important aspect of the program is that for a sample to be identified as an “unknown” sample, the value of LOI (loss on ignition in percent) should be input in the first sheet of the measured intensity file.

A novel aspect of the present work is that total 99% uncertainty can be calculated for individual datum in a given sample (treated as unknown; Table 3). This innovation if put into practice can entirely change the geochemical literature, and in fact make geochemistry a more quantitative science. Further, if an appropriate GRM is analysed as unknown and the analytical data (both mean and total uncertainly) are reported along with the field samples, the data accuracy can be statistically judged from such reports.

9. Conclusions

The XRF spectrometer calibrated under both the OLR and UWLR models clearly showed that the UWLR provides more reliable results (lower uncertainty estimates) than the OLR model commonly practiced for most XRF instruments. The sensitivity and LOD values presented for both models also supported the use of the UWLR model. The UWLR model should therefore be used routinely in such calibrations. The use of a large number of well-characterized GRMs is also recommended for this purpose as illustrated in the present work. The application of our procedure was well documented for the analysis of similarly complex rock matrices. The reporting of total uncertainty values for individual datum is highly recommended for all future geochemical research. This work for the XRF shows that such a practice is easy to achieve in any other analytical calibration procedures. As the major conclusion, we can confirm that the statistically coherent WLR model was shown to perform better than the frequently used conventional statistically incoherent OLR model.

Data Availability

The list of all compiled references (Table S1) will be available at These references are not included with the manuscript because they are too many (∼480).Similarly, as stated, the online program JSpectrom_XRFCalcUnknown will also be added onto this web portal This program needs to be available online for future use; it cannot be submitted to the journal.

Conflicts of Interest

The authors declare that they have no conflicts of interest.


This work was supported through the Newton Advanced Fellowship Award (grant NA160116) of the Royal Society, U.K., to the second author (SKV) and from the sabbatical stay of SPV at IPICYT. We are grateful to the Nanoscience and Nanotechnology National Research Laboratory (LINAN), Carbon Nanostructures and Two-Dimensional Systems Laboratory at IPICYT, and Dr. Emilio Muñoz-Sandoval for providing access to the required facilities. M. Abdelaly Rivera-Gómez is grateful to CTIC and DGAPA for a postdoctoral fellowship at the ICML-UNAM. Darío Torres-Sánchez and Mauricio Rosales-Rivera thank CONACYT for the doctoral fellowship. The GRM compilation was initiated long ago in our group by the participation of R. González-Ramírez although the bulk of the work was carried out by the present authors. We are grateful to the IER-UNAM library personnel for efficiently providing some of the literature materials for compilation and to Alfredo Quiroz-Ruiz for the maintenance of the computing facility at IER-UNAM. Diego Villanueva-López helped us during the pressed powder pellet preparation and for checking the correctness of the information in GRM databases.

Supplementary Materials

Five tables (Tables S1–S5) are provided. (Supplementary Materials)


  1. H. R. Rollinson, Using Geochemical Data: Evaluation, Presentation, Interpretation, Longman Scientific Technical, Essex, UK, 1993.
  2. J. N. Miller and J. C. Miller, Statistics and Chemometrics for Analytical Chemistry, Pearson Prentice Hall, Essex, UK, 2010.
  3. R. G. Brereton, “Chemometrics in analytical chemistry. A review,” Analyst, vol. 112, pp. 1635–1657, 1987. View at: Publisher Site | Google Scholar
  4. M. Guevara, S. P. Verma, F. Velasco-Tapia, R. Lozano-Santa Cruz, and P. Girón, “Comparison of linear regression models for quantitative geochemical analysis: an example using x-ray fluorescence spectrometry,” Geostandards and Geoanalytical Research, vol. 29, no. 3, pp. 271–284, 2005. View at: Publisher Site | Google Scholar
  5. P. R. Bevington, Data Reduction and Error Analysis for the Physical Sciences, Mc-Graw Hill Book Company, New York, NY, USA, 1969.
  6. D. York, “Least squares fitting of a straight line with correlated errors,” Earth and Planetary Science Letters, vol. 5, pp. 320–324, 1969. View at: Publisher Site | Google Scholar
  7. A. H. Kalantar, “Weighted least squares evaluation of slope from data having errors in both axes,” Trends in Analytical Chemistry, vol. 9, no. 5, pp. 149–151, 1990. View at: Publisher Site | Google Scholar
  8. K. L. Mahon, “The New “York” regression: application of an improved statistical method to geochemistry,” International Geology Review, vol. 38, no. 4, pp. 293–303, 1996. View at: Publisher Site | Google Scholar
  9. M. E. Zorn, R. D. Gibbons, and W. C. Sonzogni, “Weighted least-squares approach to calculating limits of detection and quantification by modeling variability as a function of concentration,” Analytical Chemistry, vol. 69, no. 15, pp. 3069–3075, 1997. View at: Publisher Site | Google Scholar
  10. N. R. Draper and H. Smith, Applied Regression Analysis, John Wiley & Sons, New York, NY, USA, 1998.
  11. A. Schick, “Improving weighted least-squares estimates in heteroscedastic linear regression when the variance is a function of the mean response,” Journal of Statistical Planning and Inference, vol. 76, no. 1-2, pp. 127–144, 1999. View at: Publisher Site | Google Scholar
  12. A. Sayago, M. Boccio, and A. G. Asuero, “Fitting straight lines with replicated observations by linear regression: the least squares postulates,” Critical Review of Analytical Chemistry, vol. 34, no. 1, pp. 39–50, 2004. View at: Publisher Site | Google Scholar
  13. A. Sayago and A. G. Asuero, “Fitting straight lines with replicated observations by linear regression: Part II. Testing for homogeneity of variances,” Critical Review of Analytical Chemistry, vol. 34, no. 3-4, pp. 133–146, 2004. View at: Publisher Site | Google Scholar
  14. A. G. Asuero, A. Sayago, and A. G. González, “The correlation coefficient: an overview,” Critical Review of Analytical Chemistry, vol. 36, no. 1, pp. 41–59, 2006. View at: Publisher Site | Google Scholar
  15. J. Tellinghuisen, “Weighted least-squares in calibration: what difference does it make?” Analyst, vol. 132, no. 6, pp. 536–543, 2007. View at: Publisher Site | Google Scholar
  16. S. P. Verma, L. Díaz-González, and R. González-Ramírez, “Relative efficiency of single-outlier discordancy tests for processing geochemical data on reference materials and application to instrumental calibration by a weighted least-squares linear regression model,” Geostandards and Geoanalytical Research, vol. 33, no. 1, pp. 29–49, 2009. View at: Publisher Site | Google Scholar
  17. S. P. Verma, “Geochemometrics,” Revista Mexicana de Ciencias Geológicas, vol. 29, no. 1, pp. 276–298, 2012. View at: Google Scholar
  18. S. P. Verma, Análisis Estadístico de Datos Composicionales, Universidad Nacional Autónoma de México, CDMX, Mexico, 2016.
  19. P. J. Potts, A Handbook of Silicate Rock Analysis, Blackie, Glasgow, UK, 1987.
  20. P. J. Potts and P. C. Webb, “X-ray fluorescence spectrometry,” Journal of Geochemical Exploration, vol. 44, no. 1–3, pp. 251–296, 1992. View at: Publisher Site | Google Scholar
  21. M. El Maghraoui, J.-L. Joron, J. Etoubleau, P. Cambon, and M. Treuil, “Determination of forty four major and trace elements in GPMA magmatic rock reference materials using x-ray fluorescence spectrometry (XRF) and instrumental neutron activation analysis (INAA),” Geostandards Newsletter: Journal of Geostandards and Geoanalysis, vol. 23, no. 1, pp. 59–68, 1999. View at: Publisher Site | Google Scholar
  22. K. Tani, H. Kawabata, Q. Chang, K. Sato, and Y. Tatsumi, “Quantitative analyses of silicate rock major and trace elements by X-ray fluorescence spectrometer: evaluation of analytical precision and sample preparation,” Frontiers in Research of Earth Evolution, vol. 2, pp. 1–8, 2004. View at: Google Scholar
  23. J. Enzweiler and M. A. Vendemiatto, “Analysis of sediments and soils by x-ray fluorescence spectrometry using matrix corrections based on fundamental parameters,” Geostandards and Geoanalytical Research, vol. 28, no. 1, pp. 103–112, 2005. View at: Publisher Site | Google Scholar
  24. L. P. Bédard, “Neutron activation analysis, atomic absorption and x-ray fluorescence spectrometry review for 2004-2005,” Geostandards and Geoanalytical Research, vol. 30, no. 3, pp. 183–186, 2006. View at: Publisher Site | Google Scholar
  25. K. Nakayama, Y. Shibata, and T. Nakamura, “Glass beads/x-ray fluorescence analysis of 42 components in felsic rocks,” X-Ray Spectrometry, vol. 36, no. 2, pp. 130–140, 2007. View at: Publisher Site | Google Scholar
  26. W. Wu, T. Xu, Q. Hao, Q. Wang, S. Zhang, and C. Zhao, “Applications of x-ray fluorescence analysis of rare earths in China,” Journal of the Rare Earths, vol. 28, pp. 30–36, 2010. View at: Publisher Site | Google Scholar
  27. H. Mashima, “XRF analyses of major and trace elements in silicate rocks calibrated with synthetic standard samples,” Natural Resources of Environment and Humans, vol. 6, pp. 39–50, 2016. View at: Google Scholar
  28. D. Robinson and M. C. Bennett, “XRF determination of 19 trace elements in international geochemical reference samples,” Geostandards Newsletter, vol. 5, no. 2, pp. 175–181, 1981. View at: Publisher Site | Google Scholar
  29. S. P. Verma, T. Besch, M. Guevara, and B. Schulz-Dobrich, “Determination of twelve trace elements in twenty-seven and ten major elements in twenty-three geochemical reference samples by X-Ray fluorescence spectrometry,” Geostandards Newsletter, vol. 16, no. 2, pp. 301–309, 1992. View at: Publisher Site | Google Scholar
  30. X. Wang, G. Li, Q. Zhang, and Y. Wang, “Determination of major/minor and trace elements in seamount phosphorite by XRF spectrometry,” Geostandards and Geoanalytical Research, vol. 28, no. 1, pp. 81–88, 2004. View at: Publisher Site | Google Scholar
  31. K. Govindaraju, “1987 compilation report on Ailsa Craig Granite AC-E with the participation of 128 GIT-IWG laboratories,” Geostandards Newsletter, vol. 11, no. 2, pp. 203–255, 1987. View at: Publisher Site | Google Scholar
  32. K. Govindaraju, “Report (1980) on three GIT-IWG rock reference samples: anorthosite from Greenland, AN-G; basalte d’ Essey-la-Côte, BE-N; granite de Beauvoir, MA-N,” Geostandards Newsletter, vol. 4, no. 1, pp. 49–138, 1980. View at: Publisher Site | Google Scholar
  33. E. S. Gladney and I. Roelandts, “1987 compilation of elemental concentration data for USGS BHVO-1, MAG-1, QLO-1, RGM-1, SCo-1, SDC-1, SGR-1, and STM-1,” Geostandards Newsletter, vol. 12, no. 2, pp. 253–262, 1988. View at: Publisher Site | Google Scholar
  34. K. Govindaraju, “Report (1967-1981) on four ANRT rock reference samples: diorite DR-N, serpentine UB-N, bauxite BX-N, disthene DT-N,” Geostandards Newsletter, vol. 6, no. 1, pp. 91–159, 1982. View at: Publisher Site | Google Scholar
  35. E. S. Gladney and I. Roelandts, “1987 compilation of elemental concentration data for USGS BIR-1, DNC-1 and W-2,” Geostandards Newsletter, vol. 12, no. 1, pp. 63–118, 1988. View at: Publisher Site | Google Scholar
  36. S. Abbey, C. R. McLeod, and W. Liang-Guo, “FeR-1, FeR-2, FeR-3 and FeR-4 Four Canadian iron-formation samples prepared for use as reference materials,” Tech. Rep., Geological Survey of Canada Report, Geological Survey of Canada, Ottawa, ON, Canada, 1983. View at: Google Scholar
  37. K. Govindaraju, “Report (1973-1984) in two ANRT geochemical reference samples: granite GS-N and Potash Feldspar FK-N,” Geostandards Newsletter, vol. 8, no. 2, pp. 173–206, 1984. View at: Publisher Site | Google Scholar
  38. E. S. Gladney, C. E. Burns, and I. Roelandts, “1982 compilation of elemental concentration data for the United States Geological Survey’s geochemical exploration reference samples GXR-1 to GXR-6,” Geostandards Newsletter, vol. 8, no. 2, pp. 119–154, 1984. View at: Publisher Site | Google Scholar
  39. E. S. Gladney and I. Roelandts, “1988 compilation of elemental concentration data for USGS geochemical exploration reference materials GXR-1 to GXR-6,” Geostandards Newsletter, vol. 14, no. 1, pp. 21–118, 1990. View at: Publisher Site | Google Scholar
  40. K. Govindaraju, “Report (1984) on two GIT-IWG geochemical reference samples: albite from Italy, AL-I and Iron Formation sample from Greenland, IF-G,” Geostandards Newsletter, vol. 8, no. 1, pp. 63–113, 1984. View at: Publisher Site | Google Scholar
  41. N. Imai, S. Terashima, S. Itoh, and A. Ando, “1994 compilation of analytical data for minor and trace elements in seventeen GSJ geochemical reference samples, “igneous rock series”,” Geostandards and Geoanalytical Research, vol. 19, no. 2, pp. 135–213, 1995. View at: Publisher Site | Google Scholar
  42. S. Terashima, “Elemental concentrations in nine new GSJ rock reference samples,” Geostandards Newsletter, vol. 14, no. 1, pp. 1–5, 1990. View at: Publisher Site | Google Scholar
  43. S. Terashima, S. Itoh, M. Ujiie, H. Kamioka, T. Tanaka, and H. Hattori, “Three new GSJ rock reference samples: rhyolite JR-3, gabbro JGb-2 and hornblendite JH-1,” Geostandards Newsletter, vol. 17, no. 1, pp. 1–4, 1993. View at: Publisher Site | Google Scholar
  44. S. Abbey, “Reference materials: rock samples SY-2, SY-3, MRG-1,” Tech. Rep., Energy, Mines and Resources Canada Report, Natural Resources Canada, Ottawa, ON, Canada, 1979. View at: Google Scholar
  45. K. Govindaraju, “Report (1968-1978) on two mica reference samples: biotite Mica-Fe and phlogopite Mica-Mg,” Geostandards Newsletter, vol. 3, no. 1, pp. 3–24, 1979. View at: Publisher Site | Google Scholar
  46. T. W. Steele and R. G. Hansen, “Major element data (1966-1978) for the six “Nimroc” reference samples,” Geostandards Newsletter, vol. 3, no. 2, pp. 135–172, 1979. View at: Publisher Site | Google Scholar
  47. E. S. Gladney, E. A. Jones, E. J. Nickell, and I. Roelandts, “1988 compilation of elemental concentration data for USGS DTS-1, G-1, PCC-1, and W-1,” Geostandards Newsletter, vol. 15, no. 2, pp. 199–396, 1991. View at: Publisher Site | Google Scholar
  48. V. Barnett and T. Lewis, Outliers in Statistical Data, John Wiley & Sons, Chichester, UK, 1994.
  49. K. Hayes, A. Kinsella, and N. Coffey, “A note on the use of outlier criteria in Ontario laboratory quality control schemes,” Clinical Biochemistry, vol. 40, no. 3-4, pp. 147–152, 2007. View at: Publisher Site | Google Scholar
  50. S. P. Verma, L. Díaz-González, J. A. Pérez-Garza, and M. Rosales-Rivera, “Quality control in geochemistry from a comparison of four central tendency and five dispersion estimators and example of a geochemical reference material,” Arabian Journal of Geosciences, vol. 9, p. 740, 2016. View at: Publisher Site | Google Scholar
  51. S. P. Verma, L. Díaz-González, J. A. Pérez-Garza, and M. Rosales-Rivera, “Erratum to: quality control in geochemistry from a comparison of four central tendency and five dispersion estimators and example of a geochemical reference material,” Arabian Journal of Geosciences, vol. 10, p. 24, 2017. View at: Publisher Site | Google Scholar
  52. S. P. Verma, M. Rosales-Rivera, L. Díaz-González, and A. Quiroz-Ruiz, “Improved composition of Hawaiian basalt BHVO-1 from the application of two new and three conventional recursive discordancy tests,” Turkish Journal of Earth Science, vol. 26, no. 5, pp. 331–353, 2017. View at: Publisher Site | Google Scholar
  53. S. P. Verma, R. Cruz-Huicochea, and L. Díaz-González, “Univariate data analysis system: deciphering mean compositions of island and continental arc magmas, and influence of underlying crust,” International Geology Review, vol. 55, no. 15, pp. 1922–1940, 2013. View at: Publisher Site | Google Scholar
  54. M. Rosales Rivera, Desarrollo de herramientas estadísticas computacionales con nuevos valores críticos generados por simulación computacional, Universidad Autónoma del Estado de Morelos, Cuernavaca, Morelos, Mexico, 2018.
  55. K. P. Jochum, U. Weis, B. Schwager et al., “Reference values following ISO guidelines for frequently requested rock reference materials,” Geostandards and Geoanalytical Research, vol. 40, no. 3, pp. 333–350, 2016. View at: Publisher Site | Google Scholar
  56. F. Velasco-Tapia, M. Guevara, and S. P. Verma, “Evaluation of concentration data in geochemical reference materials,” Chemie der Erde, vol. 61, no. 1, pp. 69–91, 2001. View at: Google Scholar
  57. B. Rosner, “On the detection of many outliers,” Technometrics, vol. 17, no. 2, pp. 221–227, 1975. View at: Publisher Site | Google Scholar
  58. F. E. Grubbs and G. Beck, “Extension of sample sizes and percentage points for significance tests of outlying observations,” Technometrics, vol. 14, no. 4, pp. 847–854, 1972. View at: Publisher Site | Google Scholar
  59. R. B. Jain and L. A. Pingel, “On the robustness of recursive outlier detection procedures to nonnormality,” Communications in Statistics - Theory and Methods, vol. 10, no. 13, pp. 1323–1334, 1981. View at: Publisher Site | Google Scholar
  60. R. B. Jain, “Detecting outliers: power and some other considerations,” Communications in Statistics - Theory and Methods, vol. 10, no. 22, pp. 2299–2314, 1981. View at: Publisher Site | Google Scholar
  61. M. Rosales-Rivera, L. Díaz-González, and S. P. Verma, “A new online computer program (BiDASys) for ordinary and uncertainty weighted least-squares linear regressions: case studies from food chemistry,” Revista Mexicana de Ingeniería Química, vol. 17, no. 2, pp. 507–522, 2018. View at: Publisher Site | Google Scholar
  62. J. B. Willet and J. D. Singer, “Another cautionary note about R2: its use in weighted least-squares regression analysis,” American Statistician, vol. 42, no. 3, pp. 236–238, 1988. View at: Publisher Site | Google Scholar
  63. R. M. Rousseau, “Corrections for matrix effects in X-ray fluorescence analysis—a tutorial,” Spectrochimica Acta Part B: Atomic Spectroscopy, vol. 61, no. 7, pp. 759–777, 2006. View at: Publisher Site | Google Scholar
  64. T. Shiraiwa and N. Fujino, “Theoretical calculation of fluorescent x-ray intensities in fluorescent x-ray spectrochemical analysis,” Japanese Journal of Applied Physics, vol. 5, no. 10, pp. 886–899, 1966. View at: Publisher Site | Google Scholar
  65. R. Rousseau and F. Claisse, “Theoretical alpha coefficients for the Claisse-Quintin relation for x-ray spectrochemical analysis,” X-Ray Spectrometry, vol. 3, no. 1, pp. 31–36, 1974. View at: Publisher Site | Google Scholar
  66. R. M. Rousseau, “Fundamental algorithm between concentration and intensity in XRF analysis 2-Practical application,” X-Ray Spectrometry, vol. 13, no. 3, pp. 121–125, 1984. View at: Publisher Site | Google Scholar
  67. R. M. Rousseau, J. P. Willis, and A. R. Duncan, “Practical XRF calibration procedures for major and trace elements,” X-Ray Spectrometry, vol. 25, no. 4, pp. 179–189, 1996. View at: Publisher Site | Google Scholar
  68. A. J. Klimasara, “XRF analysis—theory, experiment, and regression,” in Proceedings of X-Ray Conference (DXC) on Applications of X-Ray Analysis, Denver, CO, USA, August 1997. View at: Google Scholar
  69. B. Tan and W. Sun, “Correction method for the matrix effect in x-ray fluorescence spectrometric analysis,” X-Ray Spectrometry, vol. 27, no. 2, pp. 95–104, 1998. View at: Publisher Site | Google Scholar
  70. B. I. Kitov, “Calculation features of the fundamental parameter method in XRF,” X-Ray Spectrometry, vol. 29, no. 4, pp. 285–290, 2000. View at: Publisher Site | Google Scholar
  71. A. J. Klimasara, “Logical steps in the automated Lachamp–Triall XRF matrix correction method utilizing an electronic spreadsheet,” in Advances in XRF Analysis, vol. 42, pp. 53–83, JCPDS-International Centre for Diffraction Data, Newtown Square, PA, USA, 2000. View at: Google Scholar
  72. R. M. Rousseau, “Correction for long-term instrumental drift,” X-Ray Spectrom, vol. 31, no. 6, pp. 401–407, 2002. View at: Publisher Site | Google Scholar
  73. G. R. Lachance and R. J. Traill, “A practical solution to the matrix problem in X-ray analysis,” Canadian Journal of Spectroscopy, vol. 11, pp. 43–48, 1966. View at: Google Scholar
  74. J. P. Willis and G. R. Lachance, “Comparison between some common influence coefficient algorithms,” X-Ray Spectrometry, vol. 33, no. 3, pp. 181–188, 2004. View at: Publisher Site | Google Scholar
  75. J. P. Willis and G. R. Lachance, “A new approach to correcting theoretical emitted intensities for absorption and enhancement effects,” X-Ray Spectrometry, vol. 33, no. 3, pp. 204–211, 2004. View at: Publisher Site | Google Scholar
  76. J. W. Hosterman and F. J. Flanagan, “USGS reference samples Attapulgite ATT-1 and Bentonite CSB-1,” Geostandards and Geoanalytical Research, vol. 11, no. 1, pp. 1–9, 1987. View at: Publisher Site | Google Scholar
  77. K. Govindaraju, “1995 working values with confidence limits for twenty-six CRPG, ANRT and IWG-GIT geostandards,” Geostandards and Geoanalytical Research, vol. 19, pp. 1–32, 1995. View at: Publisher Site | Google Scholar
  78. F. J. Flanagan, “Rock reference samples, san marcos gabbro, GSM-1 and lakeview mountain tonalite, TLM-1,” Geostandards and Geoanalytical Research, vol. 10, no. 2, pp. 111–119, 1986. View at: Publisher Site | Google Scholar

Copyright © 2018 Surendra P. Verma et al. This is an open access article distributed under the Creative Commons Attribution License, which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited.

More related articles

 PDF Download Citation Citation
 Download other formatsMore
 Order printed copiesOrder

Related articles