- Email: [email protected]
On the quantitative phase analysis and amorphous content of triacylglycerols materials by X-ray Rietveld method
Accepted Manuscript Title: On the quantitative phase analysis and amorphous content of triacylglycerols materials by X-ray rietveld method Authors: Guilherme A. Calligaris, Thais L.T. da Silva, Ana Paula B. Ribeiro, Adenilson O. dos Santos, Lisandro P. Cardoso PII: DOI: Reference:
S0009-3084(17)30296-7 https://doi.org/10.1016/j.chemphyslip.2018.01.003 CPL 4629
To appear in:
Chemistry and Physics of Lipids
Received date: Revised date: Accepted date:
30-10-2017 19-12-2017 8-1-2018
Please cite this article as: Calligaris, Guilherme A., da Silva, Thais L.T., Ribeiro, Ana Paula B., dos Santos, Adenilson O., Cardoso, Lisandro P., On the quantitative phase analysis and amorphous content of triacylglycerols materials by X-ray rietveld method.Chemistry and Physics of Lipids https://doi.org/10.1016/j.chemphyslip.2018.01.003 This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to our customers we are providing this early version of the manuscript. The manuscript will undergo copyediting, typesetting, and review of the resulting proof before it is published in its final form. Please note that during the production process errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain.
On the Quantitative Phase Analysis and Amorphous Content of Triacylglycerols Materials by X-ray Rietveld Method Guilherme A. Calligaris,a Thais L. T. da Silva,b Ana Paula B. Ribeiro,b Adenilson O. dos Santosc and Lisandro P. Cardoso*a Institute of Physics Gleb Wataghin, University of Campinas – UNICAMP, 777 Sérgio Buarque de Holanda St., 13083-859 Campinas, SP, Brazil. b Department of Food Technology, Faculty of Food Engineering – UNICAMP, Cidade Universitária Zeferino Vaz, 13083-862, Campinas-SP, Brazil c CCSST, Federal University of Maranhão – UFMA, Urbano Santos St., 65900-410 Imperatriz, MA, Brazil. * Phone: +55 19 3521-5308, FAX: +55 19 3521-5376 E-mail: [email protected]
SC R
IP T
a
TE D
M
A
N
U
GraphicalAbstract
CC
EP
Graphical Abstract. Mixtures (M) and blended (B) hardfats samples involving fully hydrogenated of soybean (FHSO) and palm (FHPO) oils were characterized by XRD associated with the Rietveld structure refinement method, aiming a quantitative analysis of TAG polymorphs. The M50:50 sample was prepared to 50% of FHSO (β-form) and 50% of FHPO (β’-form). Quantitative Phase Analysis (QPA) based on Rietveld Method applied to this sample provided the expected value for them, validating the applicability of this approach in this kind of materials. After thermal treatment and melting, the QPA of the blended B50:50 sample has shown a distinct concentration of the β and β’ polymorphic forms, with the predominance of the β’-form due to the seeding process
A
Highlights:
Rietveld Method supports XRD quantitative phase analysis of polymorphism on industrial and academic appealing samples; Rietveld analysis of modified β-SSS and β’-PPS structures to allow β and β’ polymorphs content from TAG materials; Assessment of amorphous/crystalline cocoa butter content based on XRD pattern by Rietveld Method; This approach allows for improvement on the TAG polymorphic characterization compared to labeling peak intensities;
This is a LPCM – IFGW – UNICAMP manuscript
| 1
On the Quantitative Phase Analysis and Amorphous Content of Triacylglycerols Materials by X-ray Rietveld Method
SC R
IP T
Abstract. The characterization of fat components becomes very useful for formulation of shortening, margarines and fat products due to their unique properties of plasticity, texture, solubility, and aeration. However, X-ray diffraction experiments on such materials are usually limited to a qualitative evaluation of the polymorphic properties based only on the characteristic d-spacing peak intensities. In this work, interesting results based on the Rietveld Method have supported both a Quantitative Phase Analysis and Degree of Crystallinity study on industrial and academic appealing samples, such as triacylglycerol standards, fully hydrogenated vegetable oils (hardfats) and cocoa butter. This useful approach to the area of oils and fats can provide valuable information about the polymorphism and its relationship to the application of lipid materials in food science and technology. Here, the discrimination between β and β’ polymorphs on samples made of mixtures or blended hardfats was attained, and the results have shown a relevant contrast in comparison to a purely qualitative approach. Assessment of amorphous content on cocoa butter samples was achieved by isolating its contribution from the total X-ray diffraction background via mathematical tools during the whole pattern fitting.
N
U
Abbreviations
1
TE D
M
A
TAG – Triacylglycerol; XRD – X-Ray Diffraction; FHSO – Fully Hydrogenated Soybean Oil; FHPO – Fully Hydrogenated Palm Oil; CB – Cocoa Butter; CBS – Cocoa Butter Substitute; CSD – Cambridge Structural Database; RM – Rietveld Method; Rwp – Weighted Profile Rfactor; QPA – Quantitative Phase Analysis; DoC – Degree of Crystallinity.
Keywords: Triacylglycerol, Rietveld Method, Quantitative Analysis, Polymorphism, Crystallinity, X-ray Diffraction
EP
1. Introduction
CC
Long chain compounds, such as fatty acids and triacylglycerols (TAGs), can occur in differentiated crystalline forms. Three specific types of sub-cell predominate in the lipids, referring to the α, β’ and β polymorphs according to current polymorphic nomenclature (Marangoni, 2004). Generally, TAG
A
first crystallizes in the α and β’ forms, although the β form is the more stable one. The polymorphic transition is an irreversible process from the less stable to the more stable form, depending on the temperature and time involved, as well on the TAG homogeneity degree. Fats with low triacylglycerol variability quickly transform themselves into the stable β form, whilst fats with a random triacylglycerol distribution may present the β’ form indefinitely. Furthermore, factors such as
This is a LPCM – IFGW – UNICAMP manuscript
| 2
On the Quantitative Phase Analysis and Amorphous Content of Triacylglycerols Materials by X-ray Rietveld Method
formulation, cooling rate, the heat of crystallization and level of agitation, can affect the number and type of crystals formed. Nevertheless, since fats are complex mixtures of TAGs, it is possible that different polymorphic forms and liquid oil can coexist at a given temperature (Ribeiro et al., 2015). The crystalline structure of fats is important in the formulation of shortenings, margarines, spreads, chocolate and fat products in general, since each crystalline form presents unique properties concerning plasticity, texture, solubility, and aeration. Fats in the β’ form present greater functionality since they are softer and supports good aeration and creaming properties. On the other hand, the β
IP T
polymorphic form tends to produce large granular crystals, leading to sandy products with a low aeration potential, and hence compromising the macroscopic properties of the foods. Thus, β’ form is the polymorph of interest for fat-rich foods production, such as margarines, bakery and
SC R
confectionary products. Conversely, the production of chocolates requires the polymorph as the preferential crystalline phase (Ribeiro et al., 2009b).
Meanwhile, fully hydrogenated vegetable oils, also known as hardfats, has been presented as a
U
choice for lower cost and high potential usage in lipid technology. Hardfats are currently used as a
N
lipid raw material for obtaining interesterified fats and different lipid blends for food application. Besides that, hardfats from a specific oil source shows a unique and distinguished TAG profile even
A
being entirely composed of trisaturated TAGs. Therefore, these components may be regarded as
M
model systems for TAG and have been the subject of recent studies focused on modification of the crystallization process. Furthermore, soybean oil and palm oil stand out in the world vegetable oil
TE D
consumption mainly because of its nutritional quality, permanent supply, considerable economic value and high functionality, making them particularly attractive raw materials for producing hardfats for several purposes (Ribeiro et al., 2013b).
Over the years, X-ray diffraction (XRD) (or diffraction by other means like electrons, neutrons)
EP
has become a key technique to analyze crystalline materials and has mainly been used in different knowledge areas. The reason why is well-described by the International Union of Crystallography
CC
(IUCr): ‘… by “crystal” we mean any solid having an essentially discrete diffraction diagram…’ (IUCr, 1992). Regarding food science, XRD is frequently used in polymorphic studies of TAG
A
materials (Minato et al., 1997, Sonoda et al., 2004, Ribeiro et al., 2009a), providing qualitative information based only on the XRD peak intensities. Fig. 1 highlights the main d-spacing characteristic peaks of β (4.6Å) and β’ (4.2Å) polymorphs in a general TAG XRD measurement. From here, labelling the peak intensities with “s” and “m”, short for “strong” and “medium” respectively, became habitual. At this point, it is logical to judge that β polymorph concentration overcomes the β’ form in this sample since its peak intensity ratio is 60:40 (β:β’). However, such assumption is not quite true for most of the cases, as it will be shown later on, and in fact, it motivates the present work. This is a LPCM – IFGW – UNICAMP manuscript
| 3
On the Quantitative Phase Analysis and Amorphous Content of Triacylglycerols Materials by X-ray Rietveld Method
In this sense, the analysis of the XRD results (such as Fig. 1) can be further improved when Rietveld Method (RM) is applied. This method was initially developed for structures refinement for neutron diffraction (Rietveld, 1967, 1969) and then tuned for X-ray diffraction (Khattak & Cox, 1977, Malmros & Thomas, 1977, Young et al., 1977). However, one of the significant attributes of the RM is the ability to perform the whole pattern fitting, and by doing so, it can overcome some common issues on the analysis of an XRD diagram such as overlapped peaks and preferred orientation. This approach opened the possibility to precisely quantifies the crystalline phases that are presented on a
IP T
given sample, and then, it has been widely employed in distinct knowledge areas and diverse materials (Bish & Post, 1993, Srodon et al., 2001, Scrivener et al., 2004). Specifically on TAG materials, several structures have already been resolved using X-ray powder diffraction associated
SC R
with RM. Some of these contributions can be quoted as van Mechelen et al. (2006a, 2006b, 2007, 2008a, 2008b, 2008c). Moreover, the data using a RM itself can provide useful TAG model just by yielding parameters of use for the fatty materials in question, such as in Peschar et al. (2004) in which
U
the crystal structure for Cocoa Butter β-V phase was developed by using the β-SOS crystal structure as a starting point.
N
In this work, the association of XRD and RM has been applied to the analysis of TAG materials
A
with the primary objective to present a quantitative approach in food science regarding the β and
M
β’polymorphs. Also, a study on cocoa butter crystallization was used as a case for studying to obtain a quantitative analysis of the amorphous content via RM. In both scenarios, the samples were chosen
TE D
to represent typical cases found in food science and engineering literature, and thus it can establish a substantial improvement on TAG materials crystallization research.
EP
2. Materials and Methods
Samples from distinct sources were employed in order to perform an accurate quantitative study that
CC
could be tested and thus fully understood its inherent features. Commercial standards were used as a starting point on the presenting approach since its components are, by far, much more stables than
A
the observed on a usual TAG material. After that, mixtures of fully hydrogenated oils were submitted to thermal treatment, representing a typical scenario in the food industry. Finally, cocoa butter samples were prepared with cocoa butter substitute which changes its crystallinity.
2.1. Standards: LS-samples Trilaurin (LLL) – purity: 98% from Sigma-Aldrich Co. – and Tristearin (SSS) – purity: 99% from NuChek Prep Inc. – were employed here as TAG standard materials. Seven samples were prepared This is a LPCM – IFGW – UNICAMP manuscript
| 4
On the Quantitative Phase Analysis and Amorphous Content of Triacylglycerols Materials by X-ray Rietveld Method
following a linear trend in the mass ratio of 100:0, 80:20, 60:40, 50:50, 40:60, 20:80 and 0:100 between LLL:SSS, and thus were called LS-samples. The measured mass ratio of LLL and SSS is shown in Table 1.
2.2. Fully Hydrogenated Oils Fully Hydrogenated Soybean Oil (FHSO) – Cargill Agrícola S/A – and Fully Hydrogenated Palm Oil (FHPO) – SGS Agricultura e Indústria Ltda – both with at least 12 months of age were used as raw
IP T
material for the samples preparation. The analysis of their composition in TAGs was performed using a gas chromatography CGC Agilent 6850 Series GC System (Santa Clara, California, USA), with a
SC R
capillary column DB17HT Agilent (50%-phenyl-methylpolysiloxane, 15 meters in length x 0.25 mm of internal diameter and containing 0.15μm of film). The analysis conditions were: split injection, 1:100 ratio; column temperature: 250°C, programmed up to 350°C at the rate of 5°C/min; carrier gas: helium, at a low rate of 1.0 mL/min; injector temperature: 360°C; detector temperature: 375°C;
U
injected volume: 1.0 µL; sample concentration: 100 mg/5mL of tetrahydrofuran. The identification
N
of TAGs was performed by comparison of the retention times, according to the procedures of
A
Antoniosi Filho et al. (1995). The analysis was performed in triplicate for each sample. Results are
M
shown in Table 2.
Mixtures of these fully hydrogenated oils were prepared aiming mass ratio sampling of 100:0,
TE D
80:20, 60:40, 50:50, 40:60, 20:80 and 0:100 in terms of FHSO:FHPO. This original set was divided into two groups: i) M-samples: mixtures with no thermal treatment and thus preserving original state and polymorphism; ii) B-samples: mixtures in which thermal treatment of 70 ºC was employed to
EP
melt both FHSO and FHPO simultaneously until achieving visual homogeneity. After that, its recrystallization occurred under a controlled temperature of 45 ºC for 40 days in order to ensure a polymorphic stable sample, i.e., ensure a fast transition from the less stable to the most stable
CC
polymorph in the final blended sample. This whole process summarized 14 samples, M- and B-samples (mixtures and blended ones)
A
followed by the particular FHSO:FHPO ratio into its identification, as summarized in Table 1. Powder samples from both sets were prepared using a manual mill. M- and B- samples shares the same FHSO and FHPO mass ratio, which was obtained by using a precision balance (readability 0.1 mg).
2.3. Cocoa Butter and CBS In short words, Cocoa Butter (CB) plays an essential role as natural fat for chocolate formulation, however its hard access and price instability favor alternative fats such as the Cocoa Butter Equivalent This is a LPCM – IFGW – UNICAMP manuscript
| 5
On the Quantitative Phase Analysis and Amorphous Content of Triacylglycerols Materials by X-ray Rietveld Method
(CBE) and Cocoa Butter Substitutes (CBS). At the same time, crystallization and sensory characteristics from CB are quite singular and hard to achieve by using alternative fats (De Clercq et al., 2017). Here, Cocoa Butter Substitute (Fuji Oil Co.) was mixed with Brazilian Cocoa Butter (from Cargill Agricola S.A.) via a controlled temperature shaking to build the CB sample set, as shown in Table 3. The concentrations were chosen based on commonly used amounts in various products, such as the fat content in the chocolate (~32%), of which 20% comes from the naturally CB present in cocoa liquor and 12% from the added CB. The complete crystallization was ensured by performing
IP T
the tempering process according to the AOCS methodology for stabilizing confectionery fats AOCS Cd 16b-93 (AOCS, 2009). Recent work by da Silva et al. (2017) discusses crystallization properties,
SC R
the fat bloom behavior, and compatibility of such mixtures.
U
2.4. X-Ray Diffraction and Rietveld Method
N
The samples were all characterized by means of X-ray diffraction (XRD) technique. LS-, M- and Bsamples were measured on a PANalytical X’Pert MRD Pro under Bragg-Brentano geometry, using
A
4º soller slits for both incident and diffracted beams. The use of CuKα (0.15419 nm) radiation was
M
guaranteed by a pyrolytic graphite monochromator for the diffracted beam, while 4.5 to 65 degrees (2θ) scans were taken using 0.02º step size. CB-samples were studied using a Philips PW1050
TE D
goniometer with similar diffracted beam optics as abovementioned, scanned between 9 and 60 degrees (2θ) through 0.03º steps. All XRD measurements were performed at room temperature (25 ºC).
EP
Rietveld Method, in short words, provides the fit between the experimental and theoretical Xray patterns, by means of least squares iterations. It is possible to simulate a diffraction pattern and apply small modifications on the initial crystalline phases, which gives rise to the best possible fit
CC
between observed and calculated data. In order to evaluate the iterative fit process, RM introduces essential “reliability factors” (R-factors). The Rwp (weighted profile R-factor) is the most used factor
A
to mathematically show how good is the refinement. A 0% value of Rwp represents an ideally “perfect fit”, while a value of 10% can be acceptable for most cases. Rwp is known as the most sensitive evaluating-index amongst the other ones since it has an intrinsic connection with the minimized function (weighted sum of squared differences between the observed and calculated intensities). Another important R-factor is the Rexp (expected R-factor), in a few words, the best value that can be reached by Rwp, considering the initial model. The ratio of these two values (Rwp/Rexp) gives an idea of how good is the presented refinement in comparison to the best possible fit. In fact, this relationship This is a LPCM – IFGW – UNICAMP manuscript
| 6
On the Quantitative Phase Analysis and Amorphous Content of Triacylglycerols Materials by X-ray Rietveld Method
defines useful statistical parameters such as the GOF (Goodness of Fit) and χ2 by GOF2 = 2
χ2 = [Rwp ⁄Rexp ] . However, the fit evaluation only restrained to these factors is not a guarantee of a good refinement. Visual confirmation of observed and calculated data is, actually, a valid and fast way to determine the refinement quality and, consequently, the validity of the applied model and restrictions set (Toby, 2006). Meanwhile, more detailed information on the RM can be easily found in the literature (Young, 1993). In the present work, Rietveld Method is applied to allow for
IP T
Quantitative Phase Analysis (QPA) and Degree of Crystallinity (known as DoC) determination. 2.4.1. Quantitative Phase Analysis (QPA)
The employed crystalline phases were the SSS β polymorph from van Langevelde et al. (2001) and
SC R
PPS β’1-2 polymorph from van Mechelen et al. (2008c), both obtained from through Crystallographic Information Files (.cif) available via CSD (Cambridge Structural Database) v5.34 (Allen, 2002, van de Streek, 2006). The whole pattern fits were done using Fundamental Parameters approach (Cheary
U
& Coelho, 1992) available on TOPAS analytical software (Coelho et al., 2011). It is worthwhile to mention that the present work does not intend to determine TAG structures. Instead, QPA analysis
N
was done by allowing some structural parameters to vary in a limited range (restraints) from its
A
original values, in order to obtain a fully hydrogenated oil XRD pattern based only in its major single
M
TAG structure, in a similar way as Peschar et al. (2004). The employed restrictions were: a) lattice parameters - allowed to differ from the original structures by 5% at maximum; b) carbon and oxygen
TE D
atomic positions - variation up to 0.08 Å in all three lattice unit vector directions; c) hydrogen atomic positions - as heavier atoms in the unit cell provide the most significant contribution to the X-ray diffraction, the H atoms were kept constrained to their closest carbon and oxygen neighbors, i.e., its positions were linked to the C and O atoms. This was considered as the best approach to maintain the
EP
chemical bonds between the atoms along the triacylglycerol chains; and d) preferred orientation treated by spherical harmonic functions (Järvinen, 1993) to improve the fit results, avoiding non-
CC
spinner sample holder or/and by the sample textures contributions. With these restraints, physically unrealistic refined values are avoided.
A
2.4.2. Degree of Crystallinity (DoC) XRD is already used to quantify the amorphous portion of an analyzed sample by the internal standard, i.e., when a well-known crystalline material is added to the investigated sample (De La Torre et al., 2001, Martin-Marquez et al., 2009, Westphal et al., 2009, Jansen et al., 2011). However, another approach can be used based on the whole pattern fitting by RM, where the crystalline content of the sample is responsible for all the sharp XRD peaks while the amorphous content and incoherent scattering contribute to the smooth background. Ruland and Vonk (Ruland, 1961, Vonk, 1973) wellThis is a LPCM – IFGW – UNICAMP manuscript
| 7
On the Quantitative Phase Analysis and Amorphous Content of Triacylglycerols Materials by X-ray Rietveld Method
described a crystallinity determination method using XRD. The primary challenge of this method relies on the separation of the amorphous contribution from the background signal. A detailed development of this approach can be found in Riello (2004). In the present work, the amorphous content was fitted by a single Split Pseudo-Voigt (SPV) peak inserted at 2θ ~ 20.5º whilst the background followed a 3rd order polynomial Chebyshev function using TOPAS (Coelho et al., 2011). The crystalline function was described by a CB β-V (also referred as β2) polymorph structure (van Mechelen et al., 2006b). Therefore, crystalline and amorphous concentrations were obtained by their
IP T
related intensity areas over the whole pattern intensity area, in accordance with Madsen et al. (2011).
SC R
3. Results and Discussion 3.1. Quantitative Analysis (QPA)
U
3.1.1. LS-samples
All LS-samples were submitted to XRD and the results are shown in Fig. 2 where, one can see that
N
the most intense peaks from each of the standards (LLL and SSS) fall at quite similar positions (2θ:
A
19.3º, 23.05º and 24º), which turns even a typical qualitative analysis into an exhausting task.
M
However, RM uses the related crystallographic information from literature to perform the QPA, and so distinguishing contributions from LLL and SSS phases on every overlapped peak. Issues like preferred orientation can also be adequately evaluated by checking peaks from the same crystal zone
TE D
axis. The final fits are presented in Fig. 5(a), displaying the calculated intensities and their differences from the observed patterns for the LS100:0, LS50:50 and LS0:100 samples. Distinct contributions from SSS and LLL can be seen together with the LS50:50.
EP
LS-samples were solid mixtures of LLL and SSS standards, and then, the quantitative results are expected to reflect the same ratio of Table 1. The final QPA results for LS-samples are exhibited
CC
in Table 4 and the good agreement with the expected concentrations are clearly observed. Such results would be hard, or at least challenging, to be achieved only by evaluating peak intensities.
A
3.1.2. M-samples
Differently from LS-samples, M-samples are made of a mixture of FHSO and FHPO which in turn is composed by a variety of TAG species, as shown in Table 2. However, FHSO is known to be related to β polymorphic form whilst FHSO is to the β’form (Dijkstra et al., 2007, Ribeiro et al., 2013a), and then, it is reasonable to assume that M-samples behave as mixtures of “Polymorphic Standards” for β and β’, i.e., on the same way that LS-samples are mixtures of LLL and SSS.
This is a LPCM – IFGW – UNICAMP manuscript
| 8
On the Quantitative Phase Analysis and Amorphous Content of Triacylglycerols Materials by X-ray Rietveld Method
XRD was applied to distinguish from β’ and β polymorphic forms by detecting different values of d-spacing (peaks in the 2θ range) related to the contrasting β and β’ packing. For this, the most important region in the XRD is between 2θ ~15º and ~27º, which corresponds to 6 and 3 Å for dspacing values, respectively. In fact, this region is usually referred in the literature as a TAG polymorphic fingerprint region (Marangoni, 2004, Schenk & Peschar, 2004, van Mechelen et al., 2006a, van Mechelen et al., 2008a, van Mechelen et al., 2008b). Hence, β-form XRD characteristic peaks, on d-spacing values, are 4.6 Å and also the 3.86 and 3.7 Å doublet. On the other hand, β’
IP T
displays characteristic peaks for 4.2 and 3.8 Å, besides medium intensity peaks for 4.35 and 4.07 Å. All the M-samples XRD measurements are presented in Fig. 3.
As expected, M100:0 and M0:100 only shows β and β’ XRD profiles, respectively. For the
SC R
intermediary mass FHSO:FHPO ratio mixtures, XRD measurements provide a linear combination of the β and β’ contributions. The intermediary M-samples, such as M50:50 presented intensity contribution from both β and β’ characteristic peaks. The FHSO rich samples such as M80:20 and
U
M60:40 have shown β as the prevalent form, while FHPO rich samples (M20:80 and M40:60) have shown intense β’form peaks. Considering the linear mass ratio trend of M-sample set, XRD analysis
N
has confirmed a significant contribution of β polymorph for both M80:20 and M60:40 samples and
A
also, the β’ polymorph predominance for both M20:80 and M40:60 samples. This kind of intensity
M
observation as a qualitative analysis has already been employed in the literature (Ribeiro et al., 2009c, Basso et al., 2010).
TE D
At this point, the RM is employed to allow for a quantitative analysis of the M-samples. Here, the structural crystallographic information on β and β’ forms are used as initial input data represented by SSS and PPS compounds, respectively, as stated in the Materials and Methods section, and respecting the major components highlighted in Table 2. Measured XRD patterns of the M100:0 and
EP
M0:100 samples were the first to be analyzed and RM has provided a slightly modified crystalline phase for each sample, which then well represents the β (M100:0) and β’ (M0:100) forms. This is not
CC
a novel approach. Peschar et al. (2004) has already used this strategy to present a crystal structure model of the cocoa butter in the β-V phase based on a modified β-SOS. Therefore, these β and β’ phases were used as a RM starting point for the intermediary content M-samples. The refinement
A
method can quantify the polymorph fraction of the total sample mass by setting the weight that each polymorph contributes to the calculated intensity, aiming to obtain the best fit of the XRD measurements. Fig. 5(b) shows the refinement of the M100:0, M50:50 and M0:100 samples with their calculated intensities and its difference from the real measurements. β and β’ contributions to the calculated intensities are shown just below the M50:50 result.
This is a LPCM – IFGW – UNICAMP manuscript
| 9
On the Quantitative Phase Analysis and Amorphous Content of Triacylglycerols Materials by X-ray Rietveld Method
The final QPA of the M-samples Rietveld results and the corresponding Rwp values are summarized in Table 4. The obtained standard deviations of the concentration values are inside parentheses. One can observe that the β and β’ concentration values follow the same linear trend as the prepared mass ratio mixtures of FHSO:FHPO. This simple observation becomes important, since it validates the quantitative method introduced here for fully hydrogenated oils, and henceforth, it can be used with a desirable reliability. A good example of how reliable the QPA is in comparison to the qualitative visual reading of
IP T
the peak intensities is presented on the analysis of the M50:50 sample results. In fact, the M50:50 was already presented in Fig. 1, as an example of strong (s) and medium (m) intensity indexing. By only evaluating the peak intensities one can be lead to an erroneous estimative through the major
SC R
characteristic peak intensities of β and β’ forms (4.6 and 4.2 Å, respectively) that results in a ratio of 60:40. On the other hand, the QPA has provided a ratio of 52:48 which shows, with no doubt, a much better agreement with the initial FHSO:FHPO experimental mass ratio of 50.2:49.8 (Table 1).
U
3.1.3. B-samples
N
XRD measurements of the extreme B100:0 and B0:100 samples have shown similar β and β’ patterns,
A
respectively, as in the previous extreme M-samples case. However, the intermediary mass ratio Bsamples results were quite different from the M-samples, and then, the trend of β and β’
M
concentrations were no longer linear. As an example, in the XRD pattern of the B80:20 sample it has no β’ contribution, and a qualitative analysis should assume it as a pure β-form sample. Fig. 4 shows
TE D
a qualitative analysis of the B-samples through their XRD patterns. B60:40 and B50:50 samples present both β and β’ forms contributions whereas, B40:60 is a β’-form. Just for comparison purposes, the M20:80 sample pattern is also added in the figure since it represents a well-known sample which
EP
contains 20% mass ratio of β-form as determined by QPA (Table 4). A direct observation of the identified 4.6 Å characteristic peak from β-form allows observing a continuous intensity decrease in the B60:40 and B50:50 patterns, until no significant contribution of it was detected in the B40:60
CC
pattern (confirmed in the graph inset). In fact, an essentially β’ XRD pattern is observed for B40:60 sample.
A
This distinguishing behavior between M and B-samples can be explained by considering the
different sample preparation process. While M-samples were only a homogenous mixture of different mass ratio of the crystalline FHSO and FHPO, the B samples were the same FHSO:FHPO mass ratio after being melted and then recrystallized. This process is known to allow seeding feature to occur (Himawan et al., 2007, Vereecken et al., 2010) between β-SSS and β’-PPS forms from FHSO and FHPO raw materials. In this scenario, the linear behavior observed on M-samples is no longer expected to happen with the B-samples set. This is a LPCM – IFGW – UNICAMP manuscript
| 10
On the Quantitative Phase Analysis and Amorphous Content of Triacylglycerols Materials by X-ray Rietveld Method
In a similar way as presented for the LS- and M-samples, RM results of the B-samples are shown in Fig. 5(c). Observed, calculated and the difference in intensities are presented for B80:20, B50:50 and B40:60. β and β’ contributions are shown only for B50:50 sample. The fit has welladjusted the calculated pattern over the observed B-sample data, even for the weak peak at 4.6 Å which is related to the β-form concentration. Table 4 shows the concentration of β and β’ forms for all B-samples. QPA with RM of B60:40 and B50:50 delivered concentrations of 19 and 10% of the β polymorphic form, respectively. FHSO
IP T
richer samples (B100:0 and B80:20) have shown only the β form peaks, whilst B40:60, B20:80 and B0:100 only shown β’ pattern. An overall look in this table indicates the recrystallization process of B-samples favors the β’-form since the samples with intermediary concentrations has already
SC R
exhibited huge differences regarding the expected mass ratio, mainly between B80:20 and B60:40 samples.
U
It should be strengthened here an important statement considered during the QPA as to the βSSS and β’-PPS phases that were used. They represent the major structural components of the FHSO
N
and FHPO, respectively, and then, they were used in the RM to distinguish between β and β’ forms.
A
Moreover, it is well-known that the seeding process could not turn β-SSS into β’-PPS, but only, β-
M
SSS into β’-SSS. Unfortunately, as the β’-SSS crystalline phase is not yet resolved; its structure is not available in the current databases (CSD). However, it must be considered the XRD β’-form
TE D
pattern for both SSS and PPS compounds should maintain the similar characteristic profile, which presents the strongest peaks at 4.2 and 3.8 Å widely described in the literature, which, in turn, made the QPA obtained here possible.
Fig. 6(a) shows QPA results of M- and B-samples, where β’ concentration values were plotted
EP
as a function of the FHPO mass fraction. The expected β’ fraction from the weighted mass ratio during the sample preparation (Table 1) is represented as solid black squares. These squares, together
CC
with the M-samples QPA results (open red circles), are virtually lying on an ideal linear trend (dashed line). On the other hand, B-samples QPA results, marked in open blue diamonds, have shown the non-linear behavior with an enhanced β’ content, at least, above 40% of FHPO fraction for B60:40,
A
B50:50 and B40:60. The judgement on β’ fraction considering only β:β’ peak intensities ratio (as considered in Fig. 1) is also exhibited in Fig. 6(a) for comparison purposes: orange circles (Msamples) and green diamonds (B-samples). Despite the fact that β’ fraction on M100:0 and B100:0 samples was overestimated, one can easily see that the intensity ratio data have shown an overall biased behavior underestimating the β’ content, up to 12% less when compared to the expected values from sample preparation (M-samples, Table 1). This observed trend is due to their strong dependence on the background and overlapped peaks while considering the XRD intensity. This is a LPCM – IFGW – UNICAMP manuscript
| 11
On the Quantitative Phase Analysis and Amorphous Content of Triacylglycerols Materials by X-ray Rietveld Method
Whereas all QPA results in Fig. 6(a) have presented a deviation of up to 5%, the β’ fractions based on the usual peak intensity labels such as “s” (strong), “m” (medium) and “w” (weak) should only be considered a rough evaluation, since it is completely based on the human decision. Then, such approach is well affected by the “blurry” regions in between these 3 labels, as shown in Fig. 6(b) .
3.2. Crystallinity Analysis
IP T
3.2.1. CB-samples
The obtained XRD measurements from the CB-samples can be found in Fig. 7, where the β-V pattern
SC R
is observed for all samples. It is noticeable that the richer CBS samples display an overall decrease in the β-V peaks intensities whereas, a significant increase in the background intensity was only observed in the 2θ fingerprint region (14º to 26º). This result indicates that CBS has affected the CB crystallization performance, favoring the presence of amorphous content and hence, the heighten
U
background intensity.
N
RM was applied to the CB-samples to quantify the crystalline and amorphous fractions in weight for each sample. A Split Pseudo-Voigt (SPV) peak together with a smooth 3rd order
A
polynomial Chebyshev function were considered to describe the substantial increase of the
M
amorphous content and the overall background, respectively, as shown in Fig. 8. On the other hand, the observed XRD peaks well-describe the crystalline fraction during the RM.
TE D
Comparing the relative intensity areas from crystalline and amorphous fractions one can conclude that CB1 is far more crystalline than CB4, i.e., CB1 follows more closely the crystalline XRD pattern whereas, CB4 is mainly ruled by the amorphous peak, although still presenting β-V peaks. The remaining contribution for the total background is quite similar for both samples. Final
CC
EP
results are shown in Table 5, where Rwp is also exhibited.
A
4. Conclusions
In the present work, a study aiming to establish a quantitative analysis for both β and β’ triacylglycerol polymorphism and crystallinity by using X-ray diffraction associated with the Rietveld Method is presented. Rietveld Method has provided a very good fit of Trilaurin-Tristearin samples, ensuring its applicability on quantifying mixtures of TAG materials and overcoming major issues such as overlapped peaks. A much more interesting case was presented by the analysis of fully hydrogenated This is a LPCM – IFGW – UNICAMP manuscript
| 12
On the Quantitative Phase Analysis and Amorphous Content of Triacylglycerols Materials by X-ray Rietveld Method
oils mixtures (M-samples) regarding β and β’ polymorphs concentrations. Its results were well suitable in a sense that the expected linear mass ratio trend was achieved even using raw materials that are far from being as pure as the TAG standards used in the LS-samples. Furthermore, its result is a great improvement on the polymorphic characterization when compared to the widely used evaluation based on characteristic XRD peak labels intensities. Therefore, these results wellsupported the following QPA obtained from blended samples (B-samples) regarding its reliability. Furthermore, the different concentration trends observed between M- and B-samples suggest that the
IP T
present quantitative analysis was able to handle the expected seeding process in a β and β’ polymorphic system during its crystallization.
Adding CBS to Cocoa Butter samples caused an increase in the amorphous content after
SC R
tempering and complete crystallization, which was then quantified via XRD and RM. The analysis of the observed background profile allowed to accurately extract the amorphous contribution from it and then compared with the crystalline portion of the sample.
U
Summing up, the results shown here could well support research on food science and engineering, particularly in crystallization and formulation studies, since materials that are
N
industrially and/or academically appealing were employed and RM precisely defined the
A
polymorphic properties of the investigated samples. A proper discernment from polymorphs (or
M
amorphous) content can substantially improve the XRD analysis which nowadays is mainly based on peak intensities ratio or labels, i.e., as far as we know, none literature based on TAG quantitative
TE D
phase analysis by Rietveld Method was yet found. This approach can considerably enhance food science research for a more comprehensive polymorphic seeding study or even assisting chocolate formulation. Meanwhile, the present work draws attention to the benefits of using already determined structures of TAG materials from a reliable database. The existence of crystallographic phases for all
EP
polymorphic forms of the whole set of TAG components (such as Table 2) could be considered in the
CC
Rietveld analysis, and it certainly would boost the characterization of TAG materials.
A
5. Acknowledgements The authors thank the Brazilian funding agencies CNPq, CAPES, FAPEMA, and FAPESP, for the financial support of this work.
This is a LPCM – IFGW – UNICAMP manuscript
| 13
On the Quantitative Phase Analysis and Amorphous Content of Triacylglycerols Materials by X-ray Rietveld Method
References
A
CC
EP
TE D
M
A
N
U
SC R
IP T
Allen, F. H. (2002). The Cambridge Structural Database: a quarter of a million crystal structures and rising. Acta. Cryst. B 58, 380-388. Antoniosi Filho, N. R., Mendes, O. L. & Lancas, F. M. (1995). Computer-Prediction of Triacylglycerol Composition of Vegetable-Oils by Hrgc. Chromatographia 40, 557-562. AOCS (2009). Official methods and recommended practices of the AOCS, 6th ed. Urbana, Ill.: AOCS. Basso, R. C., Ribeiro, A. P. B., Masuchi, M. H., Gioielli, L. A., Gonçalves, L. A. G., dos Santos, A. O., Cardoso, L. P. & Grimaldi, R. (2010). Tripalmitin and monoacylglycerols as modifiers in the crystallisation of palm oil. Food Chem. 122, 1185-1192. Bish, D. L. & Post, J. E. (1993). Quantitative mineralogical analysis using the Rietveld full-pattern fitting method. Am. Mineral. 78, 932-940. Cheary, R. W. & Coelho, A. (1992). A fundamental parameters approach to X-ray line-profile fitting. J. Appl. Crystallogr. 25, 109-121. Coelho, A. A., Evans, J., Evans, I., Kern, A. & Parsons, S. (2011). The TOPAS symbolic computation system. Powder Diffr. 26, S22-S25. da Silva, T. L. T., Grimaldi, R., Calligaris, G. A., Cardoso, L. P. & Gonçalves, L. A. G. (2017). Crystallinity properties and crystallization behavior of chocolate fat blends. Journal of Food Science and Technology 54, 1979-1989. De Clercq, N., Kadivar, S., Van de Walle, D., De Pelsmaeker, S., Ghellynck, X. & Dewettinck, K. (2017). Functionality of cocoa butter equivalents in chocolate products. Eur. Food Res. Technol. 243, 309-321. De La Torre, A. G., Bruque, S. & Aranda, M. A. G. (2001). Rietveld quantitative amorphous content analysis. J. Appl. Crystallogr. 34, 196-202. Dijkstra, A. J., Quinn, P. J., Foubert, I., Walle, D. V. d. & Dewettinck, K. (2007). The Lipid Handbook, pp. 471-534: CRC Press. Himawan, C., MacNaughtan, W., Farhat, I. A. & Stapley, A. G. F. (2007). Polymorphic occurrence and crystallization rates of tristearin/tripalmitin mixtures under non-isothermal conditions. Eur. J. Lipid Sci. Technol. 109, 49-60. IUCr (1992). Report of the Executive Committee for 1991. Acta Crystallogr. Sect. A 48, 922-946. Jansen, D., Stabler, C., Goetz-Neunhoeffer, F., Dittrich, S. & Neubauer, J. (2011). Does Ordinary Portland Cement contain amorphous phase? A quantitative study using an external standard method. Powder Diffr. 26, 31-38. Järvinen, M. (1993). Application of symmetrized harmonics expansion to correction of the preferred orientation effect. J. Appl. Crystallogr. 26, 525-531. Khattak, C. P. & Cox, D. E. (1977). Profile analysis of X-ray powder diffractometer data: structural refinement of La0.75Sr0.25CrO3. J. Appl. Crystallogr. 10, 405-411. Madsen, I., Scarlett, N. & Kern, A. (2011). Description and survey of methodologies for the determination of amorphous content via X-ray powder diffraction. Zeitschrift für Kristallographie Crystalline Materials 226, 944-955. Malmros, G. & Thomas, J. O. (1977). Least-squares structure refinement based on profile analysis of powder film intensity data measured on an automatic microdensitometer. J. Appl. Crystallogr. 10, 7-11. Marangoni, A. (2004). Food Science and Technology, Fat Crystal Networks, pp. 1-20: CRC Press. Martin-Marquez, J., De la Torre, A. G., Aranda, M. A. G., Rincon, J. M. & Romero, M. (2009). Evolution with Temperature of Crystalline and Amorphous Phases in Porcelain Stoneware. J. Am. Ceram. Soc. 92, 229-234. Minato, A., Ueno, S., Yano, J., Smith, K., Seto, H., Amemiya, Y. & Sato, K. (1997). Thermal and structural properties of sn-1,3-dipalmitoyl-2-oleoylglycerol and sn-1,3-dioleoyl-2palmitoylglycerol binary mixtures examined with synchrotron radiation X-ray diffraction. J. Am. Oil Chem. Soc. 74, 1213-1220. This is a LPCM – IFGW – UNICAMP manuscript
| 14
On the Quantitative Phase Analysis and Amorphous Content of Triacylglycerols Materials by X-ray Rietveld Method
A
CC
EP
TE D
M
A
N
U
SC R
IP T
Peschar, R., Pop, M. M., De Ridder, D. J. A., van Mechelen, J. B., Driessen, R. A. J. & Schenk, H. (2004). Crystal structures of 1,3-distearoyl-2-oleoylglycerol and cocoa butter in the beta(V) phase reveal the driving force behind the occurrence of fat bloom on chocolate. J. Phys. Chem. B 108, 15450-15453. Ribeiro, A. P. B., Basso, R. C., dos Santos, A. O., Andrade, G. C., Cardoso, L. P. & Kieckbusch, T. G. (2013a). Hardfats as crystallization modifiers of cocoa butter. Eur. J. Lipid Sci. Technol. 115, 1462-1473. Ribeiro, A. P. B., Basso, R. C., Grimaldi, R., Gioielli, L. A., dos Santos, A. O., Cardoso, L. P. & Guaraldo Gonçalves, L. A. (2009a). Influence of chemical interesterification on thermal behavior, microstructure, polymorphism and crystallization properties of canola oil and fully hydrogenated cottonseed oil blends. Food Res. Int. 42, 1153-1162. Ribeiro, A. P. B., Basso, R. C., Grimaldi, R., Gioielli, L. A. & Gonçalves, L. A. G. (2009b). Instrumental Methods for the Evaluation of Interesterified Fats. Food Anal. Methods 2, 282302. Ribeiro, A. P. B., Basso, R. C. & Kieckbusch, T. G. (2013b). Effect of the addition of hardfats on the physical properties of cocoa butter. Eur. J. Lipid Sci. Technol. 115, 301-312. Ribeiro, A. P. B., Grimaldi, R., Gioielli, L. A., dos Santos, A. O., Cardoso, L. P. & Gonçalves, L. A. G. (2009c). Thermal Behavior, Microstructure, Polymorphism, and Crystallization Properties of Zero Trans Fats from Soybean Oil and Fully Hydrogenated Soybean Oil. Food Biophys. 4, 106-118. Ribeiro, A. P. B., Masuchi, M. H., Miyasaki, E. K., Domingues, M. A. F., Stroppa, V. L. Z., de Oliveira, G. M. & Kieckbusch, T. G. (2015). Crystallization modifiers in lipid systems. J Food Sci Tech Mys 52, 3925-3946. Riello, P. (2004). Diffraction Analysis of the Microstructure of Materials, edited by E. J. Mittemeijer & P. Scardi, pp. 167-184. Berlin, Heidelberg: Springer Berlin Heidelberg. Rietveld, H. M. (1967). Line profiles of neutron powder-diffraction peaks for structure refinement. Acta Cryst. 22, 151. Rietveld, H. M. (1969). A profile refinement method for nuclear and magnetic structures. J. Appl. Crystallogr. 2, 65-71. Ruland, W. (1961). X-Ray Determination of Crystallinity and Diffuse Disorder Scattering. Acta Cryst. 14, 1180-&. Schenk, H. & Peschar, R. (2004). Understanding the structure of chocolate. Radiat. Phys. Chem. 71, 829-835. Scrivener, K. L., Fullmann, A., Gallucci, E., Walenta, G. & Bermejo, E. (2004). Quantitative study of Portland cement hydration by X-ray diffraction/Rietveld analysis and independent methods. Cem. Concr. Res. 34, 1541-1547. Sonoda, T., Takata, Y., Ueno, S. & Sato, K. (2004). DSC and synchrotron-radiation X-ray diffraction studies on crystallization and polymorphic behavior of palm stearin in bulk and oil-in-water emulsion states. J. Am. Oil Chem. Soc. 81, 365-373. Srodon, J., Drits, V. A., McCarty, D. K., Hsieh, J. C. C. & Eberl, D. D. (2001). Quantitative X-ray diffraction analysis of clay-bearing rocks from random preparations. Clays Clay Miner. 49, 514-528. Toby, B. H. (2006). R factors in Rietveld analysis: How good is good enough? Powder Diffr. 21, 6770. van de Streek, J. (2006). Searching the Cambridge Structural Database for the `best' representative of each unique polymorph. Acta. Cryst. B 62, 567-579. van Langevelde, A., Peschar, R. & Schenk, H. (2001). Structure of [beta]-trimyristin and [beta]tristearin from high-resolution X-ray powder diffraction data. Acta. Cryst. B 57, 372-377. van Mechelen, J. B., Goubitz, K., Pop, M., Peschar, R. & Schenk, H. (2008a). Structures of monounsaturated triacylglycerols. V. The beta '(1)-2, beta '-3 and beta(2)-3 polymorphs of 1,3dilauroyl-2-oleoylglycerol (LaOLa) from synchrotron and laboratory powder diffraction data. Acta Crystallogr. Sect. B-Struct. Sci. 64, 771-779. This is a LPCM – IFGW – UNICAMP manuscript
| 15
On the Quantitative Phase Analysis and Amorphous Content of Triacylglycerols Materials by X-ray Rietveld Method
A
CC
EP
TE D
M
A
N
U
SC R
IP T
van Mechelen, J. B., Peschar, R. & Schenk, H. (2006a). Structures of mono-unsaturated triacylglycerols. I. The beta(1) polymorph. Acta Crystallogr. Sect. B-Struct. Sci. 62, 11211130. van Mechelen, J. B., Peschar, R. & Schenk, H. (2006b). Structures of mono-unsaturated triacylglycerols. II. The beta(2) polymorph. Acta Crystallogr. Sect. B-Struct. Sci. 62, 11311138. van Mechelen, J. B., Peschar, R. & Schenk, H. (2007). Structures of mono-unsaturated triacylglycerols. II. The [beta]2 polymorph. Corrigendum. Acta. Cryst. B 63, 161. van Mechelen, J. B., Peschar, R. & Schenk, H. (2008b). Structures of mono-unsaturated triacylglycerols. III. The beta-2 polymorphs of trans-mono-unsaturated triacylglycerols and related fully saturated triacylglycerols. Acta Crystallogr. Sect. B-Struct. Sci. 64, 240-248. van Mechelen, J. B., Peschar, R. & Schenk, H. (2008c). Structures of mono-unsaturated triacylglycerols. IV. The highest melting beta '-2 polymorphs of trans-mono-unsaturated triacylglycerols and related saturated TAGs and their polymorphic stability. Acta Crystallogr. Sect. B-Struct. Sci. 64, 249-259. Vereecken, J., De Graef, V., Smith, K. W., Wouters, J. & Dewettinck, K. (2010). Effect of TAG composition on the crystallization behaviour of model fat blends with the same saturated fat content. Food Res. Int. 43, 2057-2067. Vonk, C. G. (1973). Computerization of Ruland's X-Ray Method for Determination of Crystallinity in Polymers. J. Appl. Crystallogr. 6, 148-152. Westphal, T., Fullmann, T. & Pollmann, H. (2009). Rietveld quantification of amorphous portions with an internal standard-Mathematical consequences of the experimental approach. Powder Diffr. 24, 239-243. Young, R. A. (1993). The Rietveld Method. New York: Oxford University Press. Young, R. A., Mackie, P. E. & von Dreele, R. B. (1977). Application of the pattern-fitting structurerefinement method of X-ray powder diffractometer patterns. J. Appl. Crystallogr. 10, 262269.
This is a LPCM – IFGW – UNICAMP manuscript
| 16
On the Quantitative Phase Analysis and Amorphous Content of Triacylglycerols Materials by X-ray Rietveld Method
SC R
IP T
Figures
TE D
M
A
N
U
Fig. 1 | General mixture of β and β’ usually observed in food science literature. Labels like “s”, “m” and “w” are usually found in the literature to evaluate the major polymorphic form.
A
CC
EP
Fig. 2 | Observed LS-samples XRD pattern. Most of the characteristic peaks from each standard, LLL and SSS, shares very similar positions along the 2θ axis, and so even an overall content estimative based on peak intensities become impracticable.
This is a LPCM – IFGW – UNICAMP manuscript
| 17
IP T
On the Quantitative Phase Analysis and Amorphous Content of Triacylglycerols Materials by X-ray Rietveld Method
M
A
N
U
SC R
Fig. 3 | M-sample set XRD measurements. The β-form are identified by the 4.6 Å, 3.86 and 3.7 Å peaks, while β’ corresponds to the 4.2 and 3.8 Å peaks.
A
CC
EP
TE D
Fig. 4 | B-samples XRD measurements. The more intense peak from β-form is marked with an arrow in the M20:80. The inset shows in details how the intensity of the 4.6 Å characteristic peak from β-form changes.
This is a LPCM – IFGW – UNICAMP manuscript
| 18
A
N
U
SC R
IP T
On the Quantitative Phase Analysis and Amorphous Content of Triacylglycerols Materials by X-ray Rietveld Method
A
CC
EP
TE D
M
Fig. 5 | Rietveld results on fingerprint region of some of the LS-, M-, and B-samples. (a) LS100:0, LS50:50 and LS0:100 samples, in which the LLL and SSS contributions for LS50:50 are shown. RM overcome the issue of overlapped peaks from LLL and SSS. The results for M100:0, M50:50 and M0:100 samples (b) and for B80:20, B50:50 and B40:60 samples (c) are shown and the β and β’ contributions for both 50:50 mass ratio are exhibited.
Fig. 6 | Comparison between the QPA and intensity ratio results (M and B-samples) and, peak intensity labels evaluation. (a) The expected β’-form concentration from FHPO (table 1) and the ideal linear trend (traced as a guide) are plotted, together with the estimation based on peak intensity ratio. (b) β’ fraction evaluation based on peak intensity labels (“s”, “m” and “w”). “Blurry” regions represent the qualitative zones of evaluation. This is a LPCM – IFGW – UNICAMP manuscript
| 19
IP T
On the Quantitative Phase Analysis and Amorphous Content of Triacylglycerols Materials by X-ray Rietveld Method
A
CC
EP
TE D
M
A
N
U
SC R
Fig. 7 | CB-samples XRD patterns observed along the fingerprint region. Peak intensities and background level vary with the CBS addition.
Fig. 8 | Distinguishing between the crystalline and amorphous portions in CB1 (a) and CB4 (b) by using RM. A known β-V structure described the crystalline fraction whilst a Split Pseud-Voigt peak represented the amorphous contribution on the total background. The remaining background intensity follows a 3rd order polynomial Chebyshev function.
This is a LPCM – IFGW – UNICAMP manuscript
| 20
On the Quantitative Phase Analysis and Amorphous Content of Triacylglycerols Materials by X-ray Rietveld Method
Tables
FHPO [wt%] 20.0(5) 40.1(5) 49.8(5) 59.8(5) 80.0(5) 100.0
B-samples B100:0 B80:20 B60:40 B50:50 B40:60 B20:80 B0:100
U
SC R
FHSO [wt%] 100.0 80.0(5) 59.9(5) 50.2(5) 40.2(5) 20.0(5) -
A
N
M-samples M100:0 M80:20 M60:40 M50:50 M40:60 M20:80 M0:100
IP T
Table 1 | Standard TAG LS-samples contents and FHSO/FHPO mass distribution over the M- and B- samples. Deviation of 0.5 wt% is considered for all the measurements. LS-samples LLL [wt%] LLL [wt%] 100.0 LS100:0 79.9(5) 20.1(5) LS80:20 60.0(5) 40.0(5) LS60:40 50.0(5) 50.0(5) LS50:50 39.9(5) 60.1(5) LS40:60 20.0(5) 80.0(5) LS20:80 100.0 LS0:100
A
CC
EP
TE D
M
Table 2 | Composition in triacylglycerols (TAGs) of fully hydrogenated oils. Major TAGs components for each material are in bold format. P = palmitic acid; S = stearic acid. SSS = tristearoylglycerol; PSS = 1-palmitoyl-2,3-stearoyl-sn-glycerol; PPS = 1,2-palmitoyl-3- stearoyl-snglycerol; PPP = tripalmitoylglycerol. Components FHSO [%] FHPO [%] 8.19 SSS 63.79 PSS 29.81 35.89 4.67 PPS 40.80 PPP 9.39 others 1.73 5.73
This is a LPCM – IFGW – UNICAMP manuscript
| 21
On the Quantitative Phase Analysis and Amorphous Content of Triacylglycerols Materials by X-ray Rietveld Method
Table 3 | CB and CBS weight fractions in CB-samples set. Deviation of 0.5 wt% is considered for all measurements. CB-samples Id CB [ wt% ] CBS [ wt% ] 100 CB1 90.0(5) 10.0(5) CB2 85.0(5) 15.0(5) CB3 62.5(5) 37.5(5) CB4
0:100 -
100 11.81 100 6.48 0 100 6.44
A
N
U
SC R
IP T
Table 4 | Obtained LLL and SSS concentrations for the LS-samples, and also the β and β’ concentrations for M- and B-samples. Rwp values for each sample QPA. Id QPA 100:0 80:20 60:40 50:50 40:60 20:80 LLL [ wt% 100 77(4) 62(3) 44(4) 40(3) 16(5) ] LS 23(4) 38(3) 56(4) 60(3) 84(5) SSS [ wt% ] 9.79 13.13 11.88 12.37 11.46 11.45 Rwp [ % ] 100 78(2) 60(2) 52(2) 45(3) 20(3) β [ wt% ] 22(2) 40(2) 48(2) 55(3) 80(3) M β’ [ wt% ] 7.03 5.42 5.65 6.60 9.29 6.43 Rwp [ % ] 100 100 19(5) 10(5) 0 0 β [ wt% ] 0 0 81(5) 90(5) 100 100 B β’ [ wt% ] 7.23 8.24 9.10 7.61 7.54 8.29 Rwp [ % ]
A
CC
EP
TE D
M
Table 5 | Crystalline and amorphous weight fraction in the CB-samples with related Rwp values obtained by RM. CB-samples Id Crystalline [ wt% ] Amorphous [ wt% ] Rwp [ % ] 64(5) 36(5) 10.25 CB1 58(5) 42(5) 8.68 CB2 46(5) 54(5) 8.73 CB3 25(5) 75(5) 8.22 CB4
This is a LPCM – IFGW – UNICAMP manuscript
| 22
S0009-3084(17)30296-7 https://doi.org/10.1016/j.chemphyslip.2018.01.003 CPL 4629
To appear in:
Chemistry and Physics of Lipids
Received date: Revised date: Accepted date:
30-10-2017 19-12-2017 8-1-2018
Please cite this article as: Calligaris, Guilherme A., da Silva, Thais L.T., Ribeiro, Ana Paula B., dos Santos, Adenilson O., Cardoso, Lisandro P., On the quantitative phase analysis and amorphous content of triacylglycerols materials by X-ray rietveld method.Chemistry and Physics of Lipids https://doi.org/10.1016/j.chemphyslip.2018.01.003 This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to our customers we are providing this early version of the manuscript. The manuscript will undergo copyediting, typesetting, and review of the resulting proof before it is published in its final form. Please note that during the production process errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain.
On the Quantitative Phase Analysis and Amorphous Content of Triacylglycerols Materials by X-ray Rietveld Method Guilherme A. Calligaris,a Thais L. T. da Silva,b Ana Paula B. Ribeiro,b Adenilson O. dos Santosc and Lisandro P. Cardoso*a Institute of Physics Gleb Wataghin, University of Campinas – UNICAMP, 777 Sérgio Buarque de Holanda St., 13083-859 Campinas, SP, Brazil. b Department of Food Technology, Faculty of Food Engineering – UNICAMP, Cidade Universitária Zeferino Vaz, 13083-862, Campinas-SP, Brazil c CCSST, Federal University of Maranhão – UFMA, Urbano Santos St., 65900-410 Imperatriz, MA, Brazil. * Phone: +55 19 3521-5308, FAX: +55 19 3521-5376 E-mail: [email protected]
SC R
IP T
a
TE D
M
A
N
U
GraphicalAbstract
CC
EP
Graphical Abstract. Mixtures (M) and blended (B) hardfats samples involving fully hydrogenated of soybean (FHSO) and palm (FHPO) oils were characterized by XRD associated with the Rietveld structure refinement method, aiming a quantitative analysis of TAG polymorphs. The M50:50 sample was prepared to 50% of FHSO (β-form) and 50% of FHPO (β’-form). Quantitative Phase Analysis (QPA) based on Rietveld Method applied to this sample provided the expected value for them, validating the applicability of this approach in this kind of materials. After thermal treatment and melting, the QPA of the blended B50:50 sample has shown a distinct concentration of the β and β’ polymorphic forms, with the predominance of the β’-form due to the seeding process
A
Highlights:
Rietveld Method supports XRD quantitative phase analysis of polymorphism on industrial and academic appealing samples; Rietveld analysis of modified β-SSS and β’-PPS structures to allow β and β’ polymorphs content from TAG materials; Assessment of amorphous/crystalline cocoa butter content based on XRD pattern by Rietveld Method; This approach allows for improvement on the TAG polymorphic characterization compared to labeling peak intensities;
This is a LPCM – IFGW – UNICAMP manuscript
| 1
On the Quantitative Phase Analysis and Amorphous Content of Triacylglycerols Materials by X-ray Rietveld Method
SC R
IP T
Abstract. The characterization of fat components becomes very useful for formulation of shortening, margarines and fat products due to their unique properties of plasticity, texture, solubility, and aeration. However, X-ray diffraction experiments on such materials are usually limited to a qualitative evaluation of the polymorphic properties based only on the characteristic d-spacing peak intensities. In this work, interesting results based on the Rietveld Method have supported both a Quantitative Phase Analysis and Degree of Crystallinity study on industrial and academic appealing samples, such as triacylglycerol standards, fully hydrogenated vegetable oils (hardfats) and cocoa butter. This useful approach to the area of oils and fats can provide valuable information about the polymorphism and its relationship to the application of lipid materials in food science and technology. Here, the discrimination between β and β’ polymorphs on samples made of mixtures or blended hardfats was attained, and the results have shown a relevant contrast in comparison to a purely qualitative approach. Assessment of amorphous content on cocoa butter samples was achieved by isolating its contribution from the total X-ray diffraction background via mathematical tools during the whole pattern fitting.
N
U
Abbreviations
1
TE D
M
A
TAG – Triacylglycerol; XRD – X-Ray Diffraction; FHSO – Fully Hydrogenated Soybean Oil; FHPO – Fully Hydrogenated Palm Oil; CB – Cocoa Butter; CBS – Cocoa Butter Substitute; CSD – Cambridge Structural Database; RM – Rietveld Method; Rwp – Weighted Profile Rfactor; QPA – Quantitative Phase Analysis; DoC – Degree of Crystallinity.
Keywords: Triacylglycerol, Rietveld Method, Quantitative Analysis, Polymorphism, Crystallinity, X-ray Diffraction
EP
1. Introduction
CC
Long chain compounds, such as fatty acids and triacylglycerols (TAGs), can occur in differentiated crystalline forms. Three specific types of sub-cell predominate in the lipids, referring to the α, β’ and β polymorphs according to current polymorphic nomenclature (Marangoni, 2004). Generally, TAG
A
first crystallizes in the α and β’ forms, although the β form is the more stable one. The polymorphic transition is an irreversible process from the less stable to the more stable form, depending on the temperature and time involved, as well on the TAG homogeneity degree. Fats with low triacylglycerol variability quickly transform themselves into the stable β form, whilst fats with a random triacylglycerol distribution may present the β’ form indefinitely. Furthermore, factors such as
This is a LPCM – IFGW – UNICAMP manuscript
| 2
On the Quantitative Phase Analysis and Amorphous Content of Triacylglycerols Materials by X-ray Rietveld Method
formulation, cooling rate, the heat of crystallization and level of agitation, can affect the number and type of crystals formed. Nevertheless, since fats are complex mixtures of TAGs, it is possible that different polymorphic forms and liquid oil can coexist at a given temperature (Ribeiro et al., 2015). The crystalline structure of fats is important in the formulation of shortenings, margarines, spreads, chocolate and fat products in general, since each crystalline form presents unique properties concerning plasticity, texture, solubility, and aeration. Fats in the β’ form present greater functionality since they are softer and supports good aeration and creaming properties. On the other hand, the β
IP T
polymorphic form tends to produce large granular crystals, leading to sandy products with a low aeration potential, and hence compromising the macroscopic properties of the foods. Thus, β’ form is the polymorph of interest for fat-rich foods production, such as margarines, bakery and
SC R
confectionary products. Conversely, the production of chocolates requires the polymorph as the preferential crystalline phase (Ribeiro et al., 2009b).
Meanwhile, fully hydrogenated vegetable oils, also known as hardfats, has been presented as a
U
choice for lower cost and high potential usage in lipid technology. Hardfats are currently used as a
N
lipid raw material for obtaining interesterified fats and different lipid blends for food application. Besides that, hardfats from a specific oil source shows a unique and distinguished TAG profile even
A
being entirely composed of trisaturated TAGs. Therefore, these components may be regarded as
M
model systems for TAG and have been the subject of recent studies focused on modification of the crystallization process. Furthermore, soybean oil and palm oil stand out in the world vegetable oil
TE D
consumption mainly because of its nutritional quality, permanent supply, considerable economic value and high functionality, making them particularly attractive raw materials for producing hardfats for several purposes (Ribeiro et al., 2013b).
Over the years, X-ray diffraction (XRD) (or diffraction by other means like electrons, neutrons)
EP
has become a key technique to analyze crystalline materials and has mainly been used in different knowledge areas. The reason why is well-described by the International Union of Crystallography
CC
(IUCr): ‘… by “crystal” we mean any solid having an essentially discrete diffraction diagram…’ (IUCr, 1992). Regarding food science, XRD is frequently used in polymorphic studies of TAG
A
materials (Minato et al., 1997, Sonoda et al., 2004, Ribeiro et al., 2009a), providing qualitative information based only on the XRD peak intensities. Fig. 1 highlights the main d-spacing characteristic peaks of β (4.6Å) and β’ (4.2Å) polymorphs in a general TAG XRD measurement. From here, labelling the peak intensities with “s” and “m”, short for “strong” and “medium” respectively, became habitual. At this point, it is logical to judge that β polymorph concentration overcomes the β’ form in this sample since its peak intensity ratio is 60:40 (β:β’). However, such assumption is not quite true for most of the cases, as it will be shown later on, and in fact, it motivates the present work. This is a LPCM – IFGW – UNICAMP manuscript
| 3
On the Quantitative Phase Analysis and Amorphous Content of Triacylglycerols Materials by X-ray Rietveld Method
In this sense, the analysis of the XRD results (such as Fig. 1) can be further improved when Rietveld Method (RM) is applied. This method was initially developed for structures refinement for neutron diffraction (Rietveld, 1967, 1969) and then tuned for X-ray diffraction (Khattak & Cox, 1977, Malmros & Thomas, 1977, Young et al., 1977). However, one of the significant attributes of the RM is the ability to perform the whole pattern fitting, and by doing so, it can overcome some common issues on the analysis of an XRD diagram such as overlapped peaks and preferred orientation. This approach opened the possibility to precisely quantifies the crystalline phases that are presented on a
IP T
given sample, and then, it has been widely employed in distinct knowledge areas and diverse materials (Bish & Post, 1993, Srodon et al., 2001, Scrivener et al., 2004). Specifically on TAG materials, several structures have already been resolved using X-ray powder diffraction associated
SC R
with RM. Some of these contributions can be quoted as van Mechelen et al. (2006a, 2006b, 2007, 2008a, 2008b, 2008c). Moreover, the data using a RM itself can provide useful TAG model just by yielding parameters of use for the fatty materials in question, such as in Peschar et al. (2004) in which
U
the crystal structure for Cocoa Butter β-V phase was developed by using the β-SOS crystal structure as a starting point.
N
In this work, the association of XRD and RM has been applied to the analysis of TAG materials
A
with the primary objective to present a quantitative approach in food science regarding the β and
M
β’polymorphs. Also, a study on cocoa butter crystallization was used as a case for studying to obtain a quantitative analysis of the amorphous content via RM. In both scenarios, the samples were chosen
TE D
to represent typical cases found in food science and engineering literature, and thus it can establish a substantial improvement on TAG materials crystallization research.
EP
2. Materials and Methods
Samples from distinct sources were employed in order to perform an accurate quantitative study that
CC
could be tested and thus fully understood its inherent features. Commercial standards were used as a starting point on the presenting approach since its components are, by far, much more stables than
A
the observed on a usual TAG material. After that, mixtures of fully hydrogenated oils were submitted to thermal treatment, representing a typical scenario in the food industry. Finally, cocoa butter samples were prepared with cocoa butter substitute which changes its crystallinity.
2.1. Standards: LS-samples Trilaurin (LLL) – purity: 98% from Sigma-Aldrich Co. – and Tristearin (SSS) – purity: 99% from NuChek Prep Inc. – were employed here as TAG standard materials. Seven samples were prepared This is a LPCM – IFGW – UNICAMP manuscript
| 4
On the Quantitative Phase Analysis and Amorphous Content of Triacylglycerols Materials by X-ray Rietveld Method
following a linear trend in the mass ratio of 100:0, 80:20, 60:40, 50:50, 40:60, 20:80 and 0:100 between LLL:SSS, and thus were called LS-samples. The measured mass ratio of LLL and SSS is shown in Table 1.
2.2. Fully Hydrogenated Oils Fully Hydrogenated Soybean Oil (FHSO) – Cargill Agrícola S/A – and Fully Hydrogenated Palm Oil (FHPO) – SGS Agricultura e Indústria Ltda – both with at least 12 months of age were used as raw
IP T
material for the samples preparation. The analysis of their composition in TAGs was performed using a gas chromatography CGC Agilent 6850 Series GC System (Santa Clara, California, USA), with a
SC R
capillary column DB17HT Agilent (50%-phenyl-methylpolysiloxane, 15 meters in length x 0.25 mm of internal diameter and containing 0.15μm of film). The analysis conditions were: split injection, 1:100 ratio; column temperature: 250°C, programmed up to 350°C at the rate of 5°C/min; carrier gas: helium, at a low rate of 1.0 mL/min; injector temperature: 360°C; detector temperature: 375°C;
U
injected volume: 1.0 µL; sample concentration: 100 mg/5mL of tetrahydrofuran. The identification
N
of TAGs was performed by comparison of the retention times, according to the procedures of
A
Antoniosi Filho et al. (1995). The analysis was performed in triplicate for each sample. Results are
M
shown in Table 2.
Mixtures of these fully hydrogenated oils were prepared aiming mass ratio sampling of 100:0,
TE D
80:20, 60:40, 50:50, 40:60, 20:80 and 0:100 in terms of FHSO:FHPO. This original set was divided into two groups: i) M-samples: mixtures with no thermal treatment and thus preserving original state and polymorphism; ii) B-samples: mixtures in which thermal treatment of 70 ºC was employed to
EP
melt both FHSO and FHPO simultaneously until achieving visual homogeneity. After that, its recrystallization occurred under a controlled temperature of 45 ºC for 40 days in order to ensure a polymorphic stable sample, i.e., ensure a fast transition from the less stable to the most stable
CC
polymorph in the final blended sample. This whole process summarized 14 samples, M- and B-samples (mixtures and blended ones)
A
followed by the particular FHSO:FHPO ratio into its identification, as summarized in Table 1. Powder samples from both sets were prepared using a manual mill. M- and B- samples shares the same FHSO and FHPO mass ratio, which was obtained by using a precision balance (readability 0.1 mg).
2.3. Cocoa Butter and CBS In short words, Cocoa Butter (CB) plays an essential role as natural fat for chocolate formulation, however its hard access and price instability favor alternative fats such as the Cocoa Butter Equivalent This is a LPCM – IFGW – UNICAMP manuscript
| 5
On the Quantitative Phase Analysis and Amorphous Content of Triacylglycerols Materials by X-ray Rietveld Method
(CBE) and Cocoa Butter Substitutes (CBS). At the same time, crystallization and sensory characteristics from CB are quite singular and hard to achieve by using alternative fats (De Clercq et al., 2017). Here, Cocoa Butter Substitute (Fuji Oil Co.) was mixed with Brazilian Cocoa Butter (from Cargill Agricola S.A.) via a controlled temperature shaking to build the CB sample set, as shown in Table 3. The concentrations were chosen based on commonly used amounts in various products, such as the fat content in the chocolate (~32%), of which 20% comes from the naturally CB present in cocoa liquor and 12% from the added CB. The complete crystallization was ensured by performing
IP T
the tempering process according to the AOCS methodology for stabilizing confectionery fats AOCS Cd 16b-93 (AOCS, 2009). Recent work by da Silva et al. (2017) discusses crystallization properties,
SC R
the fat bloom behavior, and compatibility of such mixtures.
U
2.4. X-Ray Diffraction and Rietveld Method
N
The samples were all characterized by means of X-ray diffraction (XRD) technique. LS-, M- and Bsamples were measured on a PANalytical X’Pert MRD Pro under Bragg-Brentano geometry, using
A
4º soller slits for both incident and diffracted beams. The use of CuKα (0.15419 nm) radiation was
M
guaranteed by a pyrolytic graphite monochromator for the diffracted beam, while 4.5 to 65 degrees (2θ) scans were taken using 0.02º step size. CB-samples were studied using a Philips PW1050
TE D
goniometer with similar diffracted beam optics as abovementioned, scanned between 9 and 60 degrees (2θ) through 0.03º steps. All XRD measurements were performed at room temperature (25 ºC).
EP
Rietveld Method, in short words, provides the fit between the experimental and theoretical Xray patterns, by means of least squares iterations. It is possible to simulate a diffraction pattern and apply small modifications on the initial crystalline phases, which gives rise to the best possible fit
CC
between observed and calculated data. In order to evaluate the iterative fit process, RM introduces essential “reliability factors” (R-factors). The Rwp (weighted profile R-factor) is the most used factor
A
to mathematically show how good is the refinement. A 0% value of Rwp represents an ideally “perfect fit”, while a value of 10% can be acceptable for most cases. Rwp is known as the most sensitive evaluating-index amongst the other ones since it has an intrinsic connection with the minimized function (weighted sum of squared differences between the observed and calculated intensities). Another important R-factor is the Rexp (expected R-factor), in a few words, the best value that can be reached by Rwp, considering the initial model. The ratio of these two values (Rwp/Rexp) gives an idea of how good is the presented refinement in comparison to the best possible fit. In fact, this relationship This is a LPCM – IFGW – UNICAMP manuscript
| 6
On the Quantitative Phase Analysis and Amorphous Content of Triacylglycerols Materials by X-ray Rietveld Method
defines useful statistical parameters such as the GOF (Goodness of Fit) and χ2 by GOF2 = 2
χ2 = [Rwp ⁄Rexp ] . However, the fit evaluation only restrained to these factors is not a guarantee of a good refinement. Visual confirmation of observed and calculated data is, actually, a valid and fast way to determine the refinement quality and, consequently, the validity of the applied model and restrictions set (Toby, 2006). Meanwhile, more detailed information on the RM can be easily found in the literature (Young, 1993). In the present work, Rietveld Method is applied to allow for
IP T
Quantitative Phase Analysis (QPA) and Degree of Crystallinity (known as DoC) determination. 2.4.1. Quantitative Phase Analysis (QPA)
The employed crystalline phases were the SSS β polymorph from van Langevelde et al. (2001) and
SC R
PPS β’1-2 polymorph from van Mechelen et al. (2008c), both obtained from through Crystallographic Information Files (.cif) available via CSD (Cambridge Structural Database) v5.34 (Allen, 2002, van de Streek, 2006). The whole pattern fits were done using Fundamental Parameters approach (Cheary
U
& Coelho, 1992) available on TOPAS analytical software (Coelho et al., 2011). It is worthwhile to mention that the present work does not intend to determine TAG structures. Instead, QPA analysis
N
was done by allowing some structural parameters to vary in a limited range (restraints) from its
A
original values, in order to obtain a fully hydrogenated oil XRD pattern based only in its major single
M
TAG structure, in a similar way as Peschar et al. (2004). The employed restrictions were: a) lattice parameters - allowed to differ from the original structures by 5% at maximum; b) carbon and oxygen
TE D
atomic positions - variation up to 0.08 Å in all three lattice unit vector directions; c) hydrogen atomic positions - as heavier atoms in the unit cell provide the most significant contribution to the X-ray diffraction, the H atoms were kept constrained to their closest carbon and oxygen neighbors, i.e., its positions were linked to the C and O atoms. This was considered as the best approach to maintain the
EP
chemical bonds between the atoms along the triacylglycerol chains; and d) preferred orientation treated by spherical harmonic functions (Järvinen, 1993) to improve the fit results, avoiding non-
CC
spinner sample holder or/and by the sample textures contributions. With these restraints, physically unrealistic refined values are avoided.
A
2.4.2. Degree of Crystallinity (DoC) XRD is already used to quantify the amorphous portion of an analyzed sample by the internal standard, i.e., when a well-known crystalline material is added to the investigated sample (De La Torre et al., 2001, Martin-Marquez et al., 2009, Westphal et al., 2009, Jansen et al., 2011). However, another approach can be used based on the whole pattern fitting by RM, where the crystalline content of the sample is responsible for all the sharp XRD peaks while the amorphous content and incoherent scattering contribute to the smooth background. Ruland and Vonk (Ruland, 1961, Vonk, 1973) wellThis is a LPCM – IFGW – UNICAMP manuscript
| 7
On the Quantitative Phase Analysis and Amorphous Content of Triacylglycerols Materials by X-ray Rietveld Method
described a crystallinity determination method using XRD. The primary challenge of this method relies on the separation of the amorphous contribution from the background signal. A detailed development of this approach can be found in Riello (2004). In the present work, the amorphous content was fitted by a single Split Pseudo-Voigt (SPV) peak inserted at 2θ ~ 20.5º whilst the background followed a 3rd order polynomial Chebyshev function using TOPAS (Coelho et al., 2011). The crystalline function was described by a CB β-V (also referred as β2) polymorph structure (van Mechelen et al., 2006b). Therefore, crystalline and amorphous concentrations were obtained by their
IP T
related intensity areas over the whole pattern intensity area, in accordance with Madsen et al. (2011).
SC R
3. Results and Discussion 3.1. Quantitative Analysis (QPA)
U
3.1.1. LS-samples
All LS-samples were submitted to XRD and the results are shown in Fig. 2 where, one can see that
N
the most intense peaks from each of the standards (LLL and SSS) fall at quite similar positions (2θ:
A
19.3º, 23.05º and 24º), which turns even a typical qualitative analysis into an exhausting task.
M
However, RM uses the related crystallographic information from literature to perform the QPA, and so distinguishing contributions from LLL and SSS phases on every overlapped peak. Issues like preferred orientation can also be adequately evaluated by checking peaks from the same crystal zone
TE D
axis. The final fits are presented in Fig. 5(a), displaying the calculated intensities and their differences from the observed patterns for the LS100:0, LS50:50 and LS0:100 samples. Distinct contributions from SSS and LLL can be seen together with the LS50:50.
EP
LS-samples were solid mixtures of LLL and SSS standards, and then, the quantitative results are expected to reflect the same ratio of Table 1. The final QPA results for LS-samples are exhibited
CC
in Table 4 and the good agreement with the expected concentrations are clearly observed. Such results would be hard, or at least challenging, to be achieved only by evaluating peak intensities.
A
3.1.2. M-samples
Differently from LS-samples, M-samples are made of a mixture of FHSO and FHPO which in turn is composed by a variety of TAG species, as shown in Table 2. However, FHSO is known to be related to β polymorphic form whilst FHSO is to the β’form (Dijkstra et al., 2007, Ribeiro et al., 2013a), and then, it is reasonable to assume that M-samples behave as mixtures of “Polymorphic Standards” for β and β’, i.e., on the same way that LS-samples are mixtures of LLL and SSS.
This is a LPCM – IFGW – UNICAMP manuscript
| 8
On the Quantitative Phase Analysis and Amorphous Content of Triacylglycerols Materials by X-ray Rietveld Method
XRD was applied to distinguish from β’ and β polymorphic forms by detecting different values of d-spacing (peaks in the 2θ range) related to the contrasting β and β’ packing. For this, the most important region in the XRD is between 2θ ~15º and ~27º, which corresponds to 6 and 3 Å for dspacing values, respectively. In fact, this region is usually referred in the literature as a TAG polymorphic fingerprint region (Marangoni, 2004, Schenk & Peschar, 2004, van Mechelen et al., 2006a, van Mechelen et al., 2008a, van Mechelen et al., 2008b). Hence, β-form XRD characteristic peaks, on d-spacing values, are 4.6 Å and also the 3.86 and 3.7 Å doublet. On the other hand, β’
IP T
displays characteristic peaks for 4.2 and 3.8 Å, besides medium intensity peaks for 4.35 and 4.07 Å. All the M-samples XRD measurements are presented in Fig. 3.
As expected, M100:0 and M0:100 only shows β and β’ XRD profiles, respectively. For the
SC R
intermediary mass FHSO:FHPO ratio mixtures, XRD measurements provide a linear combination of the β and β’ contributions. The intermediary M-samples, such as M50:50 presented intensity contribution from both β and β’ characteristic peaks. The FHSO rich samples such as M80:20 and
U
M60:40 have shown β as the prevalent form, while FHPO rich samples (M20:80 and M40:60) have shown intense β’form peaks. Considering the linear mass ratio trend of M-sample set, XRD analysis
N
has confirmed a significant contribution of β polymorph for both M80:20 and M60:40 samples and
A
also, the β’ polymorph predominance for both M20:80 and M40:60 samples. This kind of intensity
M
observation as a qualitative analysis has already been employed in the literature (Ribeiro et al., 2009c, Basso et al., 2010).
TE D
At this point, the RM is employed to allow for a quantitative analysis of the M-samples. Here, the structural crystallographic information on β and β’ forms are used as initial input data represented by SSS and PPS compounds, respectively, as stated in the Materials and Methods section, and respecting the major components highlighted in Table 2. Measured XRD patterns of the M100:0 and
EP
M0:100 samples were the first to be analyzed and RM has provided a slightly modified crystalline phase for each sample, which then well represents the β (M100:0) and β’ (M0:100) forms. This is not
CC
a novel approach. Peschar et al. (2004) has already used this strategy to present a crystal structure model of the cocoa butter in the β-V phase based on a modified β-SOS. Therefore, these β and β’ phases were used as a RM starting point for the intermediary content M-samples. The refinement
A
method can quantify the polymorph fraction of the total sample mass by setting the weight that each polymorph contributes to the calculated intensity, aiming to obtain the best fit of the XRD measurements. Fig. 5(b) shows the refinement of the M100:0, M50:50 and M0:100 samples with their calculated intensities and its difference from the real measurements. β and β’ contributions to the calculated intensities are shown just below the M50:50 result.
This is a LPCM – IFGW – UNICAMP manuscript
| 9
On the Quantitative Phase Analysis and Amorphous Content of Triacylglycerols Materials by X-ray Rietveld Method
The final QPA of the M-samples Rietveld results and the corresponding Rwp values are summarized in Table 4. The obtained standard deviations of the concentration values are inside parentheses. One can observe that the β and β’ concentration values follow the same linear trend as the prepared mass ratio mixtures of FHSO:FHPO. This simple observation becomes important, since it validates the quantitative method introduced here for fully hydrogenated oils, and henceforth, it can be used with a desirable reliability. A good example of how reliable the QPA is in comparison to the qualitative visual reading of
IP T
the peak intensities is presented on the analysis of the M50:50 sample results. In fact, the M50:50 was already presented in Fig. 1, as an example of strong (s) and medium (m) intensity indexing. By only evaluating the peak intensities one can be lead to an erroneous estimative through the major
SC R
characteristic peak intensities of β and β’ forms (4.6 and 4.2 Å, respectively) that results in a ratio of 60:40. On the other hand, the QPA has provided a ratio of 52:48 which shows, with no doubt, a much better agreement with the initial FHSO:FHPO experimental mass ratio of 50.2:49.8 (Table 1).
U
3.1.3. B-samples
N
XRD measurements of the extreme B100:0 and B0:100 samples have shown similar β and β’ patterns,
A
respectively, as in the previous extreme M-samples case. However, the intermediary mass ratio Bsamples results were quite different from the M-samples, and then, the trend of β and β’
M
concentrations were no longer linear. As an example, in the XRD pattern of the B80:20 sample it has no β’ contribution, and a qualitative analysis should assume it as a pure β-form sample. Fig. 4 shows
TE D
a qualitative analysis of the B-samples through their XRD patterns. B60:40 and B50:50 samples present both β and β’ forms contributions whereas, B40:60 is a β’-form. Just for comparison purposes, the M20:80 sample pattern is also added in the figure since it represents a well-known sample which
EP
contains 20% mass ratio of β-form as determined by QPA (Table 4). A direct observation of the identified 4.6 Å characteristic peak from β-form allows observing a continuous intensity decrease in the B60:40 and B50:50 patterns, until no significant contribution of it was detected in the B40:60
CC
pattern (confirmed in the graph inset). In fact, an essentially β’ XRD pattern is observed for B40:60 sample.
A
This distinguishing behavior between M and B-samples can be explained by considering the
different sample preparation process. While M-samples were only a homogenous mixture of different mass ratio of the crystalline FHSO and FHPO, the B samples were the same FHSO:FHPO mass ratio after being melted and then recrystallized. This process is known to allow seeding feature to occur (Himawan et al., 2007, Vereecken et al., 2010) between β-SSS and β’-PPS forms from FHSO and FHPO raw materials. In this scenario, the linear behavior observed on M-samples is no longer expected to happen with the B-samples set. This is a LPCM – IFGW – UNICAMP manuscript
| 10
On the Quantitative Phase Analysis and Amorphous Content of Triacylglycerols Materials by X-ray Rietveld Method
In a similar way as presented for the LS- and M-samples, RM results of the B-samples are shown in Fig. 5(c). Observed, calculated and the difference in intensities are presented for B80:20, B50:50 and B40:60. β and β’ contributions are shown only for B50:50 sample. The fit has welladjusted the calculated pattern over the observed B-sample data, even for the weak peak at 4.6 Å which is related to the β-form concentration. Table 4 shows the concentration of β and β’ forms for all B-samples. QPA with RM of B60:40 and B50:50 delivered concentrations of 19 and 10% of the β polymorphic form, respectively. FHSO
IP T
richer samples (B100:0 and B80:20) have shown only the β form peaks, whilst B40:60, B20:80 and B0:100 only shown β’ pattern. An overall look in this table indicates the recrystallization process of B-samples favors the β’-form since the samples with intermediary concentrations has already
SC R
exhibited huge differences regarding the expected mass ratio, mainly between B80:20 and B60:40 samples.
U
It should be strengthened here an important statement considered during the QPA as to the βSSS and β’-PPS phases that were used. They represent the major structural components of the FHSO
N
and FHPO, respectively, and then, they were used in the RM to distinguish between β and β’ forms.
A
Moreover, it is well-known that the seeding process could not turn β-SSS into β’-PPS, but only, β-
M
SSS into β’-SSS. Unfortunately, as the β’-SSS crystalline phase is not yet resolved; its structure is not available in the current databases (CSD). However, it must be considered the XRD β’-form
TE D
pattern for both SSS and PPS compounds should maintain the similar characteristic profile, which presents the strongest peaks at 4.2 and 3.8 Å widely described in the literature, which, in turn, made the QPA obtained here possible.
Fig. 6(a) shows QPA results of M- and B-samples, where β’ concentration values were plotted
EP
as a function of the FHPO mass fraction. The expected β’ fraction from the weighted mass ratio during the sample preparation (Table 1) is represented as solid black squares. These squares, together
CC
with the M-samples QPA results (open red circles), are virtually lying on an ideal linear trend (dashed line). On the other hand, B-samples QPA results, marked in open blue diamonds, have shown the non-linear behavior with an enhanced β’ content, at least, above 40% of FHPO fraction for B60:40,
A
B50:50 and B40:60. The judgement on β’ fraction considering only β:β’ peak intensities ratio (as considered in Fig. 1) is also exhibited in Fig. 6(a) for comparison purposes: orange circles (Msamples) and green diamonds (B-samples). Despite the fact that β’ fraction on M100:0 and B100:0 samples was overestimated, one can easily see that the intensity ratio data have shown an overall biased behavior underestimating the β’ content, up to 12% less when compared to the expected values from sample preparation (M-samples, Table 1). This observed trend is due to their strong dependence on the background and overlapped peaks while considering the XRD intensity. This is a LPCM – IFGW – UNICAMP manuscript
| 11
On the Quantitative Phase Analysis and Amorphous Content of Triacylglycerols Materials by X-ray Rietveld Method
Whereas all QPA results in Fig. 6(a) have presented a deviation of up to 5%, the β’ fractions based on the usual peak intensity labels such as “s” (strong), “m” (medium) and “w” (weak) should only be considered a rough evaluation, since it is completely based on the human decision. Then, such approach is well affected by the “blurry” regions in between these 3 labels, as shown in Fig. 6(b) .
3.2. Crystallinity Analysis
IP T
3.2.1. CB-samples
The obtained XRD measurements from the CB-samples can be found in Fig. 7, where the β-V pattern
SC R
is observed for all samples. It is noticeable that the richer CBS samples display an overall decrease in the β-V peaks intensities whereas, a significant increase in the background intensity was only observed in the 2θ fingerprint region (14º to 26º). This result indicates that CBS has affected the CB crystallization performance, favoring the presence of amorphous content and hence, the heighten
U
background intensity.
N
RM was applied to the CB-samples to quantify the crystalline and amorphous fractions in weight for each sample. A Split Pseudo-Voigt (SPV) peak together with a smooth 3rd order
A
polynomial Chebyshev function were considered to describe the substantial increase of the
M
amorphous content and the overall background, respectively, as shown in Fig. 8. On the other hand, the observed XRD peaks well-describe the crystalline fraction during the RM.
TE D
Comparing the relative intensity areas from crystalline and amorphous fractions one can conclude that CB1 is far more crystalline than CB4, i.e., CB1 follows more closely the crystalline XRD pattern whereas, CB4 is mainly ruled by the amorphous peak, although still presenting β-V peaks. The remaining contribution for the total background is quite similar for both samples. Final
CC
EP
results are shown in Table 5, where Rwp is also exhibited.
A
4. Conclusions
In the present work, a study aiming to establish a quantitative analysis for both β and β’ triacylglycerol polymorphism and crystallinity by using X-ray diffraction associated with the Rietveld Method is presented. Rietveld Method has provided a very good fit of Trilaurin-Tristearin samples, ensuring its applicability on quantifying mixtures of TAG materials and overcoming major issues such as overlapped peaks. A much more interesting case was presented by the analysis of fully hydrogenated This is a LPCM – IFGW – UNICAMP manuscript
| 12
On the Quantitative Phase Analysis and Amorphous Content of Triacylglycerols Materials by X-ray Rietveld Method
oils mixtures (M-samples) regarding β and β’ polymorphs concentrations. Its results were well suitable in a sense that the expected linear mass ratio trend was achieved even using raw materials that are far from being as pure as the TAG standards used in the LS-samples. Furthermore, its result is a great improvement on the polymorphic characterization when compared to the widely used evaluation based on characteristic XRD peak labels intensities. Therefore, these results wellsupported the following QPA obtained from blended samples (B-samples) regarding its reliability. Furthermore, the different concentration trends observed between M- and B-samples suggest that the
IP T
present quantitative analysis was able to handle the expected seeding process in a β and β’ polymorphic system during its crystallization.
Adding CBS to Cocoa Butter samples caused an increase in the amorphous content after
SC R
tempering and complete crystallization, which was then quantified via XRD and RM. The analysis of the observed background profile allowed to accurately extract the amorphous contribution from it and then compared with the crystalline portion of the sample.
U
Summing up, the results shown here could well support research on food science and engineering, particularly in crystallization and formulation studies, since materials that are
N
industrially and/or academically appealing were employed and RM precisely defined the
A
polymorphic properties of the investigated samples. A proper discernment from polymorphs (or
M
amorphous) content can substantially improve the XRD analysis which nowadays is mainly based on peak intensities ratio or labels, i.e., as far as we know, none literature based on TAG quantitative
TE D
phase analysis by Rietveld Method was yet found. This approach can considerably enhance food science research for a more comprehensive polymorphic seeding study or even assisting chocolate formulation. Meanwhile, the present work draws attention to the benefits of using already determined structures of TAG materials from a reliable database. The existence of crystallographic phases for all
EP
polymorphic forms of the whole set of TAG components (such as Table 2) could be considered in the
CC
Rietveld analysis, and it certainly would boost the characterization of TAG materials.
A
5. Acknowledgements The authors thank the Brazilian funding agencies CNPq, CAPES, FAPEMA, and FAPESP, for the financial support of this work.
This is a LPCM – IFGW – UNICAMP manuscript
| 13
On the Quantitative Phase Analysis and Amorphous Content of Triacylglycerols Materials by X-ray Rietveld Method
References
A
CC
EP
TE D
M
A
N
U
SC R
IP T
Allen, F. H. (2002). The Cambridge Structural Database: a quarter of a million crystal structures and rising. Acta. Cryst. B 58, 380-388. Antoniosi Filho, N. R., Mendes, O. L. & Lancas, F. M. (1995). Computer-Prediction of Triacylglycerol Composition of Vegetable-Oils by Hrgc. Chromatographia 40, 557-562. AOCS (2009). Official methods and recommended practices of the AOCS, 6th ed. Urbana, Ill.: AOCS. Basso, R. C., Ribeiro, A. P. B., Masuchi, M. H., Gioielli, L. A., Gonçalves, L. A. G., dos Santos, A. O., Cardoso, L. P. & Grimaldi, R. (2010). Tripalmitin and monoacylglycerols as modifiers in the crystallisation of palm oil. Food Chem. 122, 1185-1192. Bish, D. L. & Post, J. E. (1993). Quantitative mineralogical analysis using the Rietveld full-pattern fitting method. Am. Mineral. 78, 932-940. Cheary, R. W. & Coelho, A. (1992). A fundamental parameters approach to X-ray line-profile fitting. J. Appl. Crystallogr. 25, 109-121. Coelho, A. A., Evans, J., Evans, I., Kern, A. & Parsons, S. (2011). The TOPAS symbolic computation system. Powder Diffr. 26, S22-S25. da Silva, T. L. T., Grimaldi, R., Calligaris, G. A., Cardoso, L. P. & Gonçalves, L. A. G. (2017). Crystallinity properties and crystallization behavior of chocolate fat blends. Journal of Food Science and Technology 54, 1979-1989. De Clercq, N., Kadivar, S., Van de Walle, D., De Pelsmaeker, S., Ghellynck, X. & Dewettinck, K. (2017). Functionality of cocoa butter equivalents in chocolate products. Eur. Food Res. Technol. 243, 309-321. De La Torre, A. G., Bruque, S. & Aranda, M. A. G. (2001). Rietveld quantitative amorphous content analysis. J. Appl. Crystallogr. 34, 196-202. Dijkstra, A. J., Quinn, P. J., Foubert, I., Walle, D. V. d. & Dewettinck, K. (2007). The Lipid Handbook, pp. 471-534: CRC Press. Himawan, C., MacNaughtan, W., Farhat, I. A. & Stapley, A. G. F. (2007). Polymorphic occurrence and crystallization rates of tristearin/tripalmitin mixtures under non-isothermal conditions. Eur. J. Lipid Sci. Technol. 109, 49-60. IUCr (1992). Report of the Executive Committee for 1991. Acta Crystallogr. Sect. A 48, 922-946. Jansen, D., Stabler, C., Goetz-Neunhoeffer, F., Dittrich, S. & Neubauer, J. (2011). Does Ordinary Portland Cement contain amorphous phase? A quantitative study using an external standard method. Powder Diffr. 26, 31-38. Järvinen, M. (1993). Application of symmetrized harmonics expansion to correction of the preferred orientation effect. J. Appl. Crystallogr. 26, 525-531. Khattak, C. P. & Cox, D. E. (1977). Profile analysis of X-ray powder diffractometer data: structural refinement of La0.75Sr0.25CrO3. J. Appl. Crystallogr. 10, 405-411. Madsen, I., Scarlett, N. & Kern, A. (2011). Description and survey of methodologies for the determination of amorphous content via X-ray powder diffraction. Zeitschrift für Kristallographie Crystalline Materials 226, 944-955. Malmros, G. & Thomas, J. O. (1977). Least-squares structure refinement based on profile analysis of powder film intensity data measured on an automatic microdensitometer. J. Appl. Crystallogr. 10, 7-11. Marangoni, A. (2004). Food Science and Technology, Fat Crystal Networks, pp. 1-20: CRC Press. Martin-Marquez, J., De la Torre, A. G., Aranda, M. A. G., Rincon, J. M. & Romero, M. (2009). Evolution with Temperature of Crystalline and Amorphous Phases in Porcelain Stoneware. J. Am. Ceram. Soc. 92, 229-234. Minato, A., Ueno, S., Yano, J., Smith, K., Seto, H., Amemiya, Y. & Sato, K. (1997). Thermal and structural properties of sn-1,3-dipalmitoyl-2-oleoylglycerol and sn-1,3-dioleoyl-2palmitoylglycerol binary mixtures examined with synchrotron radiation X-ray diffraction. J. Am. Oil Chem. Soc. 74, 1213-1220. This is a LPCM – IFGW – UNICAMP manuscript
| 14
On the Quantitative Phase Analysis and Amorphous Content of Triacylglycerols Materials by X-ray Rietveld Method
A
CC
EP
TE D
M
A
N
U
SC R
IP T
Peschar, R., Pop, M. M., De Ridder, D. J. A., van Mechelen, J. B., Driessen, R. A. J. & Schenk, H. (2004). Crystal structures of 1,3-distearoyl-2-oleoylglycerol and cocoa butter in the beta(V) phase reveal the driving force behind the occurrence of fat bloom on chocolate. J. Phys. Chem. B 108, 15450-15453. Ribeiro, A. P. B., Basso, R. C., dos Santos, A. O., Andrade, G. C., Cardoso, L. P. & Kieckbusch, T. G. (2013a). Hardfats as crystallization modifiers of cocoa butter. Eur. J. Lipid Sci. Technol. 115, 1462-1473. Ribeiro, A. P. B., Basso, R. C., Grimaldi, R., Gioielli, L. A., dos Santos, A. O., Cardoso, L. P. & Guaraldo Gonçalves, L. A. (2009a). Influence of chemical interesterification on thermal behavior, microstructure, polymorphism and crystallization properties of canola oil and fully hydrogenated cottonseed oil blends. Food Res. Int. 42, 1153-1162. Ribeiro, A. P. B., Basso, R. C., Grimaldi, R., Gioielli, L. A. & Gonçalves, L. A. G. (2009b). Instrumental Methods for the Evaluation of Interesterified Fats. Food Anal. Methods 2, 282302. Ribeiro, A. P. B., Basso, R. C. & Kieckbusch, T. G. (2013b). Effect of the addition of hardfats on the physical properties of cocoa butter. Eur. J. Lipid Sci. Technol. 115, 301-312. Ribeiro, A. P. B., Grimaldi, R., Gioielli, L. A., dos Santos, A. O., Cardoso, L. P. & Gonçalves, L. A. G. (2009c). Thermal Behavior, Microstructure, Polymorphism, and Crystallization Properties of Zero Trans Fats from Soybean Oil and Fully Hydrogenated Soybean Oil. Food Biophys. 4, 106-118. Ribeiro, A. P. B., Masuchi, M. H., Miyasaki, E. K., Domingues, M. A. F., Stroppa, V. L. Z., de Oliveira, G. M. & Kieckbusch, T. G. (2015). Crystallization modifiers in lipid systems. J Food Sci Tech Mys 52, 3925-3946. Riello, P. (2004). Diffraction Analysis of the Microstructure of Materials, edited by E. J. Mittemeijer & P. Scardi, pp. 167-184. Berlin, Heidelberg: Springer Berlin Heidelberg. Rietveld, H. M. (1967). Line profiles of neutron powder-diffraction peaks for structure refinement. Acta Cryst. 22, 151. Rietveld, H. M. (1969). A profile refinement method for nuclear and magnetic structures. J. Appl. Crystallogr. 2, 65-71. Ruland, W. (1961). X-Ray Determination of Crystallinity and Diffuse Disorder Scattering. Acta Cryst. 14, 1180-&. Schenk, H. & Peschar, R. (2004). Understanding the structure of chocolate. Radiat. Phys. Chem. 71, 829-835. Scrivener, K. L., Fullmann, A., Gallucci, E., Walenta, G. & Bermejo, E. (2004). Quantitative study of Portland cement hydration by X-ray diffraction/Rietveld analysis and independent methods. Cem. Concr. Res. 34, 1541-1547. Sonoda, T., Takata, Y., Ueno, S. & Sato, K. (2004). DSC and synchrotron-radiation X-ray diffraction studies on crystallization and polymorphic behavior of palm stearin in bulk and oil-in-water emulsion states. J. Am. Oil Chem. Soc. 81, 365-373. Srodon, J., Drits, V. A., McCarty, D. K., Hsieh, J. C. C. & Eberl, D. D. (2001). Quantitative X-ray diffraction analysis of clay-bearing rocks from random preparations. Clays Clay Miner. 49, 514-528. Toby, B. H. (2006). R factors in Rietveld analysis: How good is good enough? Powder Diffr. 21, 6770. van de Streek, J. (2006). Searching the Cambridge Structural Database for the `best' representative of each unique polymorph. Acta. Cryst. B 62, 567-579. van Langevelde, A., Peschar, R. & Schenk, H. (2001). Structure of [beta]-trimyristin and [beta]tristearin from high-resolution X-ray powder diffraction data. Acta. Cryst. B 57, 372-377. van Mechelen, J. B., Goubitz, K., Pop, M., Peschar, R. & Schenk, H. (2008a). Structures of monounsaturated triacylglycerols. V. The beta '(1)-2, beta '-3 and beta(2)-3 polymorphs of 1,3dilauroyl-2-oleoylglycerol (LaOLa) from synchrotron and laboratory powder diffraction data. Acta Crystallogr. Sect. B-Struct. Sci. 64, 771-779. This is a LPCM – IFGW – UNICAMP manuscript
| 15
On the Quantitative Phase Analysis and Amorphous Content of Triacylglycerols Materials by X-ray Rietveld Method
A
CC
EP
TE D
M
A
N
U
SC R
IP T
van Mechelen, J. B., Peschar, R. & Schenk, H. (2006a). Structures of mono-unsaturated triacylglycerols. I. The beta(1) polymorph. Acta Crystallogr. Sect. B-Struct. Sci. 62, 11211130. van Mechelen, J. B., Peschar, R. & Schenk, H. (2006b). Structures of mono-unsaturated triacylglycerols. II. The beta(2) polymorph. Acta Crystallogr. Sect. B-Struct. Sci. 62, 11311138. van Mechelen, J. B., Peschar, R. & Schenk, H. (2007). Structures of mono-unsaturated triacylglycerols. II. The [beta]2 polymorph. Corrigendum. Acta. Cryst. B 63, 161. van Mechelen, J. B., Peschar, R. & Schenk, H. (2008b). Structures of mono-unsaturated triacylglycerols. III. The beta-2 polymorphs of trans-mono-unsaturated triacylglycerols and related fully saturated triacylglycerols. Acta Crystallogr. Sect. B-Struct. Sci. 64, 240-248. van Mechelen, J. B., Peschar, R. & Schenk, H. (2008c). Structures of mono-unsaturated triacylglycerols. IV. The highest melting beta '-2 polymorphs of trans-mono-unsaturated triacylglycerols and related saturated TAGs and their polymorphic stability. Acta Crystallogr. Sect. B-Struct. Sci. 64, 249-259. Vereecken, J., De Graef, V., Smith, K. W., Wouters, J. & Dewettinck, K. (2010). Effect of TAG composition on the crystallization behaviour of model fat blends with the same saturated fat content. Food Res. Int. 43, 2057-2067. Vonk, C. G. (1973). Computerization of Ruland's X-Ray Method for Determination of Crystallinity in Polymers. J. Appl. Crystallogr. 6, 148-152. Westphal, T., Fullmann, T. & Pollmann, H. (2009). Rietveld quantification of amorphous portions with an internal standard-Mathematical consequences of the experimental approach. Powder Diffr. 24, 239-243. Young, R. A. (1993). The Rietveld Method. New York: Oxford University Press. Young, R. A., Mackie, P. E. & von Dreele, R. B. (1977). Application of the pattern-fitting structurerefinement method of X-ray powder diffractometer patterns. J. Appl. Crystallogr. 10, 262269.
This is a LPCM – IFGW – UNICAMP manuscript
| 16
On the Quantitative Phase Analysis and Amorphous Content of Triacylglycerols Materials by X-ray Rietveld Method
SC R
IP T
Figures
TE D
M
A
N
U
Fig. 1 | General mixture of β and β’ usually observed in food science literature. Labels like “s”, “m” and “w” are usually found in the literature to evaluate the major polymorphic form.
A
CC
EP
Fig. 2 | Observed LS-samples XRD pattern. Most of the characteristic peaks from each standard, LLL and SSS, shares very similar positions along the 2θ axis, and so even an overall content estimative based on peak intensities become impracticable.
This is a LPCM – IFGW – UNICAMP manuscript
| 17
IP T
On the Quantitative Phase Analysis and Amorphous Content of Triacylglycerols Materials by X-ray Rietveld Method
M
A
N
U
SC R
Fig. 3 | M-sample set XRD measurements. The β-form are identified by the 4.6 Å, 3.86 and 3.7 Å peaks, while β’ corresponds to the 4.2 and 3.8 Å peaks.
A
CC
EP
TE D
Fig. 4 | B-samples XRD measurements. The more intense peak from β-form is marked with an arrow in the M20:80. The inset shows in details how the intensity of the 4.6 Å characteristic peak from β-form changes.
This is a LPCM – IFGW – UNICAMP manuscript
| 18
A
N
U
SC R
IP T
On the Quantitative Phase Analysis and Amorphous Content of Triacylglycerols Materials by X-ray Rietveld Method
A
CC
EP
TE D
M
Fig. 5 | Rietveld results on fingerprint region of some of the LS-, M-, and B-samples. (a) LS100:0, LS50:50 and LS0:100 samples, in which the LLL and SSS contributions for LS50:50 are shown. RM overcome the issue of overlapped peaks from LLL and SSS. The results for M100:0, M50:50 and M0:100 samples (b) and for B80:20, B50:50 and B40:60 samples (c) are shown and the β and β’ contributions for both 50:50 mass ratio are exhibited.
Fig. 6 | Comparison between the QPA and intensity ratio results (M and B-samples) and, peak intensity labels evaluation. (a) The expected β’-form concentration from FHPO (table 1) and the ideal linear trend (traced as a guide) are plotted, together with the estimation based on peak intensity ratio. (b) β’ fraction evaluation based on peak intensity labels (“s”, “m” and “w”). “Blurry” regions represent the qualitative zones of evaluation. This is a LPCM – IFGW – UNICAMP manuscript
| 19
IP T
On the Quantitative Phase Analysis and Amorphous Content of Triacylglycerols Materials by X-ray Rietveld Method
A
CC
EP
TE D
M
A
N
U
SC R
Fig. 7 | CB-samples XRD patterns observed along the fingerprint region. Peak intensities and background level vary with the CBS addition.
Fig. 8 | Distinguishing between the crystalline and amorphous portions in CB1 (a) and CB4 (b) by using RM. A known β-V structure described the crystalline fraction whilst a Split Pseud-Voigt peak represented the amorphous contribution on the total background. The remaining background intensity follows a 3rd order polynomial Chebyshev function.
This is a LPCM – IFGW – UNICAMP manuscript
| 20
On the Quantitative Phase Analysis and Amorphous Content of Triacylglycerols Materials by X-ray Rietveld Method
Tables
FHPO [wt%] 20.0(5) 40.1(5) 49.8(5) 59.8(5) 80.0(5) 100.0
B-samples B100:0 B80:20 B60:40 B50:50 B40:60 B20:80 B0:100
U
SC R
FHSO [wt%] 100.0 80.0(5) 59.9(5) 50.2(5) 40.2(5) 20.0(5) -
A
N
M-samples M100:0 M80:20 M60:40 M50:50 M40:60 M20:80 M0:100
IP T
Table 1 | Standard TAG LS-samples contents and FHSO/FHPO mass distribution over the M- and B- samples. Deviation of 0.5 wt% is considered for all the measurements. LS-samples LLL [wt%] LLL [wt%] 100.0 LS100:0 79.9(5) 20.1(5) LS80:20 60.0(5) 40.0(5) LS60:40 50.0(5) 50.0(5) LS50:50 39.9(5) 60.1(5) LS40:60 20.0(5) 80.0(5) LS20:80 100.0 LS0:100
A
CC
EP
TE D
M
Table 2 | Composition in triacylglycerols (TAGs) of fully hydrogenated oils. Major TAGs components for each material are in bold format. P = palmitic acid; S = stearic acid. SSS = tristearoylglycerol; PSS = 1-palmitoyl-2,3-stearoyl-sn-glycerol; PPS = 1,2-palmitoyl-3- stearoyl-snglycerol; PPP = tripalmitoylglycerol. Components FHSO [%] FHPO [%] 8.19 SSS 63.79 PSS 29.81 35.89 4.67 PPS 40.80 PPP 9.39 others 1.73 5.73
This is a LPCM – IFGW – UNICAMP manuscript
| 21
On the Quantitative Phase Analysis and Amorphous Content of Triacylglycerols Materials by X-ray Rietveld Method
Table 3 | CB and CBS weight fractions in CB-samples set. Deviation of 0.5 wt% is considered for all measurements. CB-samples Id CB [ wt% ] CBS [ wt% ] 100 CB1 90.0(5) 10.0(5) CB2 85.0(5) 15.0(5) CB3 62.5(5) 37.5(5) CB4
0:100 -
100 11.81 100 6.48 0 100 6.44
A
N
U
SC R
IP T
Table 4 | Obtained LLL and SSS concentrations for the LS-samples, and also the β and β’ concentrations for M- and B-samples. Rwp values for each sample QPA. Id QPA 100:0 80:20 60:40 50:50 40:60 20:80 LLL [ wt% 100 77(4) 62(3) 44(4) 40(3) 16(5) ] LS 23(4) 38(3) 56(4) 60(3) 84(5) SSS [ wt% ] 9.79 13.13 11.88 12.37 11.46 11.45 Rwp [ % ] 100 78(2) 60(2) 52(2) 45(3) 20(3) β [ wt% ] 22(2) 40(2) 48(2) 55(3) 80(3) M β’ [ wt% ] 7.03 5.42 5.65 6.60 9.29 6.43 Rwp [ % ] 100 100 19(5) 10(5) 0 0 β [ wt% ] 0 0 81(5) 90(5) 100 100 B β’ [ wt% ] 7.23 8.24 9.10 7.61 7.54 8.29 Rwp [ % ]
A
CC
EP
TE D
M
Table 5 | Crystalline and amorphous weight fraction in the CB-samples with related Rwp values obtained by RM. CB-samples Id Crystalline [ wt% ] Amorphous [ wt% ] Rwp [ % ] 64(5) 36(5) 10.25 CB1 58(5) 42(5) 8.68 CB2 46(5) 54(5) 8.73 CB3 25(5) 75(5) 8.22 CB4
This is a LPCM – IFGW – UNICAMP manuscript
| 22