Quantitative Assessment of Triacylglycerol Crystallite Thickness by 1H

Subscriber access provided by University of Newcastle, Australia

Article

Quantitative assessment of triacylglycerol crystallite thickness by 1H spin-diffusion NMR Adrian Voda, Ruud Den Adel, Kees van Malssen, and John van Duynhoven Cryst. Growth Des., Just Accepted Manuscript • DOI: 10.1021/acs.cgd.6b00501 • Publication Date (Web): 15 Mar 2017 Downloaded from http://pubs.acs.org on March 16, 2017

Just Accepted “Just Accepted” manuscripts have been peer-reviewed and accepted for publication. They are posted online prior to technical editing, formatting for publication and author proofing. The American Chemical Society provides “Just Accepted” as a free service to the research community to expedite the dissemination of scientific material as soon as possible after acceptance. “Just Accepted” manuscripts appear in full in PDF format accompanied by an HTML abstract. “Just Accepted” manuscripts have been fully peer reviewed, but should not be considered the official version of record. They are accessible to all readers and citable by the Digital Object Identifier (DOI®). “Just Accepted” is an optional service offered to authors. Therefore, the “Just Accepted” Web site may not include all articles that will be published in the journal. After a manuscript is technically edited and formatted, it will be removed from the “Just Accepted” Web site and published as an ASAP article. Note that technical editing may introduce minor changes to the manuscript text and/or graphics which could affect content, and all legal disclaimers and ethical guidelines that apply to the journal pertain. ACS cannot be held responsible for errors or consequences arising from the use of information contained in these “Just Accepted” manuscripts.

Crystal Growth & Design is published by the American Chemical Society. 1155 Sixteenth Street N.W., Washington, DC 20036 Published by American Chemical Society. Copyright © American Chemical Society. However, no copyright claim is made to original U.S. Government works, or works produced by employees of any Commonwealth realm Crown government in the course of their duties.

Page 1 of 34

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Crystal Growth & Design

Quantitative assessment of triacylglycerol crystallite thickness by 1H spin-diffusion NMR Adrian Voda *,1, Ruud den Adel 1, Kees van Malssen 1, John van Duynhoven 1,2 1

Unilever R&D Vlaardingen, Olivier van Noortlaan 120, 3130 AC Vlaardingen, The Netherlands

2

Laboratory of Biophysics and Wageningen NMR Centre, Wageningen University, Dreijenlaan 3, 6703 HA Wageningen, The Netherlands

ABSTRACT: Properties and consumer appreciation of fat containing food products are greatly influenced by the growth of the fat crystalline domains and their morphology. Various analytical methods can be exploited to provide insight into the multi-length scale microstructure of fat networks, but very few are able to discriminate structural features between different ways of processing fats. We selected in this work methods that highlight the mesostructure of fat crystalline systems. In this respect, 1H spin-diffusion NMR and XRD methods were employed to determine the fat crystallite thicknesses (crystalline domain size). This is the first attempt to quantify the mesostructure domains of triglyceride-based fat crystals by means of 1H spin-diffusion NMR experiments. The crystalline domain size determined by spin-diffusion NMR were found to be in good agreement with the crystallite thickness determined by the Scherrer analysis of the first order diffraction line from SAXS data. These results demonstrate the ability of the NMR technique to characterize the mesostructure of fats in a quantitative manner. This method is of particular interest for the structure analysis of fats, especially because of the possibility to quantify the size of the crystalline domains in diluted systems where scattering techniques struggle with the amount of diffracting material.

*

[email protected], +31 10 4606708, Unilever R&D Vlaardingen, Olivier van Noortlaan 120, 3130 AC Vlaardingen, The Netherlands

ACS Paragon Plus Environment

1


1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Page 2 of 34

Quantitative assessment of triacylglycerol crystallite thickness by 1H spin-diffusion NMR Adrian Voda *,1, Ruud den Adel 1, Kees van Malssen 1, John van Duynhoven 1,2 1

Unilever R&D Vlaardingen, Olivier van Noortlaan 120, 3130 AC Vlaardingen, The

Netherlands 2

Laboratory of Biophysics and Wageningen NMR Centre, Wageningen University, Dreijenlaan

3, 6703 HA Wageningen, The Netherlands

*

To whom correspondence should be addressed: [email protected]

KEYWORDS: triglyceride, triacylglycerol, fat, crystal network, domain size, crystallite thickness, mesostructure, Nuclear Magnetic Resonance (NMR), spin-diffusion, small angle Xray scattering (SAXS), X-ray diffraction (XRD). ABSTRACT: Properties and consumer appreciation of fat containing food products are greatly influenced by the growth of the fat crystalline domains and their morphology. Various analytical methods can be exploited to provide insight into the multi-length scale microstructure of fat networks, but very few are able to discriminate structural features between different ways of processing fats. We selected in this work methods that highlight the mesostructure of fat crystalline systems. In this respect, 1H spin-diffusion NMR and XRD methods were employed to


2

Page 3 of 34

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60


determine the fat crystallite thicknesses (crystalline domain size). This is the first attempt to quantify the mesostructure domains of triglyceride-based fat crystals by means of 1H spindiffusion NMR experiments. The crystalline domain size determined by spin-diffusion NMR for two fat blends with different composition crystallised at different cooling rates were found in the range of 30 – 64 nm (blend A) and 28 – 40 nm (blend B). This results are in good agreement with the crystallite thickness determined by the Scherrer analysis of the first order diffraction line from SAXS data, which were found in the range of 35 – 70 nm (blend A) and 29 – 51 nm (blend B). These results demonstrate the ability of the NMR technique to characterize the mesostructure of fats in a quantitative manner. This method is of particular interest for the structure analysis of fats, especially because of the possibility to quantify the size of the crystalline domains in diluted systems where scattering techniques struggle with the amount of diffracting material.


3


1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Page 4 of 34

INTRODUCTION Vegetable fats predominantly consist of triacylglycerols (TAGs), which at room temperature exist as a mixture of crystals and liquid oil. The fat crystals can form a continuous network, thereby giving firmness to the system, while the oil exists in a continuous liquid phase.1 Many factors affect the formation and the properties of crystal networks in food systems. TAG molecules can pack in different crystalline polymorphs (α, β’, and β), which exhibit significantly different melting temperatures.2-4 Structural properties of TAG crystals are influenced by molecular properties of TAG such as saturation/unsaturation of the fatty acid moieties, glycerol conformations, symmetry/asymmetry of the fatty acid compositions connected to the glycerol groups and so forth.4-7 The phase-behaviour of lipid components is known to have a major impact on the formation of fat crystal network properties, which have a multi-length scale architecture (Figure 1). These multi-length scale crystal networks can structure significant amounts of liquid oil8, and determine consumer-relevant physical properties of fat-continuous products like margarines9 and chocolate, such as melting and solidification, texture, and density.6 Features of the multi-length scale crystal network can be assessed by a multitude of analytical techniques, each of them playing a role at a specific length scale. A representation of the hierarchical multi-length scale architecture of fats10 is depicted in Figure 1. The morphology of the building blocks of a crystalline fat system, from the TAG molecules up to the macrostructure of the fat network was recently reviewed in extensive detail.10-13 However, when trying to quantify structural parameters of crystalline and non-crystalline fat phases, many techniques struggle with the lack of contrast required to discriminate these phases. So far, the analysis at the nanoscale has been well represented by Wide and Small Angle X-ray scattering (WAXS and SAXS).10,14 FT-IR and solid-state NMR techniques were also involved in revealing


4

Page 5 of 34

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60


nanostructure information.10,15 The accessible length scale remains limited, which is sufficient only for estimating periodical spacings within TAG bilayers and crystallite thickness (domain size) up to a value limited by the SAXS capability of the instrument. In the case of synchrotron radiation this can be as high as 200 nm. Moreover, in crystalline-diluted systems the sensitivity of X-ray diffraction methods becomes critical. However, the other two dimensions of a fat crystallite, the length and the width, much larger than the thickness, are too large to access by SAXS. A series of recent studies13,16-18 demonstrated the benefit of ultra-small angle X-ray scattering (USAXS) to quantify long-range repeating structures in fats with sizes on the micrometer range and to obtain information about the morphology of fat aggregates and fractal dimensions. To retrieve information on larger architectural entities like crystal stacks and clusters to confirm the scattering data, electron microscopy could be useful but requires invasive sample preparation procedures in order to visualize crystalline structures by removing the liquid fat. Different attempts to remove the oil while trying to preserve the crystalline entities involved the use of detergent aqueous solutions19-21 or organic solvents that dissolve the liquid fat mainly.22-24 These methods have a high risk of introducing artifacts. Crystal platelets of model TAG blends could be visualized in this manner and their dimensions quantified by cryoTEM.10,23-25 So far only examples of crystallite thickness estimation in model TAGs exist by cryo-TEM and the versatility of the cold solvent extraction in practical-use fats has not been demonstrated yet. In the tens of micrometer range and above, crystalline structures can well be investigated by various microscopy and spectroscopy techniques. A relatively underexplored range remains a part of the mesostructure, from few tens up to few hundred nanometers, whereas already mentioned, X-ray diffraction and TEM methods experience limitations. A technique that specifically can provide the thickness of the crystalline domains is 1H spin-diffusion NMR.26,27


5


1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Page 6 of 34

The method was successfully applied to synthetic copolymers, polymer blends and it is generally applicable to phase-heterogeneous polymer materials.27-38 Proton spin-diffusion correlates structure and molecular mobility information with the heterogeneously distributed phase domains. This is a unique feature that allows one to study various systems without limitations by imaging contrast (optical or electron), scattering density or periodicity. The purpose of this work is to investigate the applicability of 1H spin-diffusion NMR to quantify the thickness of the fat crystallites in a fat network. Two different fat systems were considered in this study, consisting of blends of TAGs with different fatty acids composition and diversity. The fat blends were subject to different crystallization rates by applying either a very fast or a moderate-slow cooling from the melt, thus crystallites growth and size are expected to be different.39,40 Proton spin-diffusion method was applied in combination with a doublequantum (DQ) dipolar filter to select the magnetization of the most rigid/crystalline phase. Other magnetization filters can also be employed to select either the low- or the high-molecular mobility spin populations.30,35,37 To the best of our knowledge this is the first attempt to estimate the thickness of TAG crystallites in fat systems by 1H spin-diffusion NMR, after we presented the idea at the “NMR in Foods” conference in 2012.41 Comparison of NMR results with crystallite thickness determined by Scherrer analysis of the first order diffraction peak from SAXS data is discussed.


6

Page 7 of 34

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60


Figure 1 Hierarchical multi-length scale structure representation of a crystalline fat system.

METHODOLOGY Materials. Two different TAG blends were investigated and for simplicity we will denote these as “blend A” and “blend B”. Blend A is an interesterified fat blend with a melting point of 45 °C, while blend B is a hydrogenated fat with a melting point of 38 °C. The fatty acid composition determined by Fatty Acid Methyl Esthers analysis (FAME) is denoted in Table 1. Table 1 Fatty acid composition of TAG blends used in this work. Fatty acid

Blend A

Blend B

Stearic acid; St (C18:0)

5%

13%

Palmitic acid; P (C16:0)

50%

8%

Myristic acid; M (C14:0)

7%

14%


7


1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Page 8 of 34

Lauric acid; L (C12:0)

20%

50%

Unsaturated; U (C18:x)

12%

6%

Others

6%

9%

The fat blends were molten at 80 °C and crystallization was performed following two cooling routes: 1) a moderate-low cooling controlled at a rate of 5 °/min down to 20 °C, after that the sample was stored at 5 °C; 2) a very fast cooling that was achieved by plunging drops of molten fat in liquid nitrogen (-196 °C), after that the sample was stored as well at 5 °C. All samples, were kept at 5 °C for a day in order to allow the crystal polymorph to gain the β’ configuration, as confirmed by WAXS measurements. NMR, WAXS and SAXS measurements were carried out at 5 °C. 1

H spin-diffusion NMR. Spin-diffusion is a process that describes the migration of nuclear

magnetization through space.27 This phenomenon is mediated by the dipole-dipole interaction between neighboring nuclear spins. In solid matter, the dipolar transfer of magnetization depends on the spin species and abundance, and on the strength of the dipolar couplings with respect to other interactions. In solids, where most of the alike spins are dipolar coupled, it can be assumed that every nuclear dipole transfers per unit time a fraction of its magnetization to the neighbors, and receives as well magnetization from them. Hence, the homonuclear transfer of (longitudinal) magnetization is a space and time dependent process that can be described by the Fick’s second law of diffusion:

= ∇

(1)


8

Page 9 of 34

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60


where Mz is the longitudinal magnetization and D is the spin-diffusion coefficient. In solids that exhibit microphase separation the process of spin-diffusion can be exploited to retrieve the size of the morphological domains. In order to determine domain sizes by the 1H spin-diffusion experiment, one needs to create a gradient in magnetization across domains with different molecular mobility. Certain magnetization filters can be applied that allow for a selection of magnetization in one of the domain type (rigid or mobile), hence generating the magnetization gradient required. The time dependence of the diffusion of magnetization contains information about the size of the domains between which the transfer occurs. Values for domain sizes from about 0.5 nm up to few hundred nm have been reported in the literature. In triglyceride systems in solid phase, dipolar couplings between 1H pairs of the akyl chains of the fatty acids can provide the means of the spin-diffusion process. The morphology of fat crystallites is known to be lamellar, with a length and width an order of magnitude larger than the thickness. Therefore, a one-dimensional morphology with a repeating unit schematically represented in Figure 2 is assumed.


9


1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Page 10 of 34

Figure 2 Repeating unit for 1D spin-diffusion in lamellar morphology for two- and three-phase domains. The domain sizes for the rigid and mobile phases are d R and d M , respectively, and dI for the interface. We assume that spin-diffusion takes place from a source R (rigid) with low molecular mobility represented by the crystalline TAGs into a sink M (mobile) with larger mobility corresponding to the non-crystalline phase of the TAGs. The solutions of the diffusion equation (1) for a one-dimensional spin-diffusion case in a two-component system (lamellar morphology in Figure 2) can be determined analytically (SI1).35,42,43 When crystallizing a TAG saturated/unsaturated mixture, a two-component system will be formed. Crystallites, also referred to as rigid domains, are formed by stacks of bi-layers of TAG molecules in an environment of non-crystalline, less ordered TAGs, forming the mobile domain.


10

Page 11 of 34

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60


NMR experiments. Proton wideline NMR spectra, DQ buildup curves and spin-diffusion data were measured on a Bruker DSX-300 spectrometer operating at 300 MHz Larmor frequency for 1

H. A solenoidal rf coil with a dead time of the order of 5 µs was used to carry out the

experiments. The length of the π/2 pulse was set to 1.5 µs, the dwell time was 2 µs, and the recycle delay was set to 20 s for all measurements. 1H spin-diffusion measurements were conducted using a generic scheme consisting of a double-quantum (DQ) dipolar filter, a spindiffusion period, and an acquisition period as presented in Figure 3. The magnetization gradient was created by a dipolar filter that excites DQ coherences (Figure 3) and which mainly selects the magnetization of the rigid phase.35 The value of the excitation/reconversion time, τ, was set at 9 µs corresponding to the rising region of the DQ buildup curve for each sample.

Figure 3 Scheme for the spin-diffusion experiment with a DQ filter. The first 90°-180°-90° block excites DQ coherences that evolve for a short time tDQ. These coherences are converted by the following block into z-magnetization. The spin-diffusion takes place during the time interval of duration td. The experimental wideline spectra were decomposed in three components using MestReNova package. A Gaussian lineshape was assigned to the broad spectral component


11


1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Page 12 of 34

corresponding to the low molecular mobility fraction. The highest mobility fraction was fitted with a Lorentzian whilst the intermediate fraction was considered as a liniar combination of Gaussian and Lorentzian lineshapes. The relative phase fractions obtained by lineshape deconvolution as a function of spindiffusion time were used to determine the domain sizes. The data were treated according to spindiffusion across a three-phase model (rigid, interface and mobile). A simplification to a twophase model (Figure 2) was also considered, where the interface amount was evenly distributed between the rigid and mobile fractions. The spin-diffusion curves (phase fractions as a function of square root of the diffusion time) were simulated based on the solutions of one-dimensional diffusion process in lamellar morphology, from a finite source (rigid) to a finite sink (interface, mobile). Simulation parameters include the spin-diffusion coefficient, the proton density and the domain size of each phase fraction.

SAXS experiments. Small Angle X-ray diffraction (SAXS) analyses were performed on a Bruker D8-Discover diffractometer. In a θ/θ configuration the long spacings (thickness of the repeating bi-layers) of the fat crystallites were measured. In order to obtain information about the thickness of the fat crystallites the Full Width at Half Maximum (FWHM) value of the first order diffraction line was used. The average crystallite thickness, d, is calculated using the Scherrer formula:

=

(2)


12

Page 13 of 34

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60


where θ and λ are the Bragg angle and X-ray wavelength (in nanometer), respectively. β is the line width of the peak in radians corrected for instrumental broadening. K and λ are 0.9

24

and

0.15418 nm, respectively. The Scherrer equation shows an inverse relationship between crystallite size and peak width: the broader the peak, the smaller the crystallites. The instrumental line broadening is 0.180° 2-θ, which is determined using the FWHM of the 1 1 1 Si reflection of NIST standard reference material 640. The samples were measured at 5 °C using a temperature controlled stage. FWHM was calculated using the EVA software from Bruker AXS.

RESULTS AND DISCUSSION Proton NMR spectra. 1H NMR spectra of the fat blend A together with its lineshape deconvolution are presented in Figure 4 (the spectrum of blend B is similar). The spectra of all four samples have been subject to multi-component lineshape deconvolution. The best fit yielded the least residuals in the case of three components reflecting three phase fractions with different 1

H molecular mobility. The broad line described by a Gaussian shape was attributed to the most

rigid fraction of the fat systems corresponding to the crystalline entities defining the repeating bilayers of TAGs. The narrow lineshape was described by a Lorentzian and it is attributed to the TAGs that undergo the highest chain mobility and form the mobile phase. The third component reveals an intermediate lineshape and it was described as a combination of Gaussian and Lorentzian that can be interpreted as a semi-rigid, as well as a semi-mobile phase, that can also be regarded as an interface between the crystalline and the mobile domains. The presence of the intermediate-mobility interface was confirmed by additional heteronuclear

1

H-13C WISE

experiments (SI2). The phase fractions obtained by spectral deconvolution are presented in


13


1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Page 14 of 34

Figure 5 and the corresponding linewidths at half intensity in Table 2. As expected, a large amount of crystalline fraction is revealed in all samples. The line width (Table 2) of the mobile fraction indicates fast molecular dynamics of the protons, which corresponds to a liquid phase. Certain differences between phase fraction and molecular dynamics can be observed in Figure 5 and Table 2 regarding the result of the crystallization rate. Fast crystallization facilitates the formation of an intermediate mobility fraction, which builds up primarily on the expense of the crystalline phase. A smaller crystalline fraction due to faster cooling can be explained by a forced embedding of less fitting TAGs in the crystal lattice, hence resulting in crystal errors with increased molecular mobility. Molecular motions of the protons in the crystalline phase remain about the same irrespective of the cooling rate of both fat blends. Differences between molecular mobility of the intermediate phase are significant between the two fat blends and as a function of crystallization rate for blend B. The intermediate phase of blend A consists of TAGs with higher chain mobility than blend B (Table 1), which indicates a less hindered molecular motion. The mobility of the intermediate phase of blend B is higher for faster cooling. A significant lower mobility is observed in the mobile fraction upon fast crystallization. This can be due to the forced entrapment of certain TAGs within the solid structure, which at slow cooling remain “free”, thus reducing their chain mobility.


14

Page 15 of 34

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60


Figure 4 1H NMR spectrum of TAG blend A with the corresponding spectral line deconvolution by means of Gaussian and Lorentzian line-shapes. The broad and narrow components correspond to respectively crystalline and mobile fractions of the fat network, whilst the intermediate linewidth component represents the semi-rigid/semi-mobile interface.


15


1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Page 16 of 34

a

Blend A

b

Blend B


16

Page 17 of 34

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60


Figure 5 Phase fraction composition by 1H NMR of (a) blend A and (b) blend B after slow and fast crystallization steps. Table 2 Full width at half intensity for 1H NMR spectra components of TAG blends under fast and slow crystallization conditions. phase fraction *

∆ν1/2 (kHz) **

R

39.6

M

0.3

I

11.6

R

40.5

M

0.4

I

5.5

R

39.2

M

0.9

I

10.8

R

39

M

0.9

I

7.6

Slow crystallization Blend A

Blend B

Fast crystallization Blend A

Blend B

*R = rigid, M = mobile, I = intermediate/interface **Errors are less than 5 % Double-quantum filter for magnetization. Proton spin-diffusion experiments observe the equilibration of spatially heterogeneous magnetization over the sample. To create the required magnetization gradient, Zeeman order filters can be applied that allow the magnetization of one


17


1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Page 18 of 34

of the phases to pass through the filter while the magnetization of the other phase is filtered out.30 Such filters exploit the differences in the magnetization transport properties of the spins in the phase fractions with different 1H mobility. We have employed a double quantum (DQ) filter to create the magnetization gradient between the crystallites and the mobile fraction of the fat blends. Multiple-quantum NMR is a sensitive measure of dipolar interactions in many solid and liquid materials. DQ filter can be set in such a way as to select the magnetization only from the most rigid part of a heterogeneous sample. By choosing appropriate excitation/reconversion periods (Figure 3) of the DQ filter, magnetization corresponding to the stronger dipolar couplings will pass through the filter whereas components with weaker dipolar couplings are filtered out. The value of τ can be chosen by inspecting the DQ build-up curve shown in Figure 6 for blend A, recorded using the DQ pulse sequence with a filter time td= 2 µs (Figure 3). The maximum of the DQ buildup curve appears at very short excitation/reconversion times, τ, around 10 µs for all samples. In this range of τ values, the mobile component is completely filtered out. The DQ filtered spectra for different values of τ are shown in Figure 7. A value of 9 µs excitation/reconversion time was chosen, which still keeps the filter efficiency30 close to unity with a reasonable value of signal-to-noise ratio.


18

Page 19 of 34

1.0

DQ normalized signal

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60


0.8 0.6 0.4 0.2 0.0 0

10

20

30

40

50

60

70

80

τ [µs]

Figure 6 1H DQ build-up curve of TAG blend A obtained from the data recorded using the pulse sequence in Figure 3 with tDQ = td = 2 µs.

Figure 7 1H DQ filtered spectra for different τ values measured using the pulse sequence of Figure 3 with tDQ= td=2 µs. The spectra intensities are rescaled to ease the visual inspection.


19


1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Page 20 of 34

Crystallite sizes by 1H spin-diffusion NMR and SAXS. Samples were first investigated by WAXS and all TAG blends revealed a β’ crystal polymorph. Subsequently, SAXS measurements were carried out and an example of the first order diffraction line is presented in Figure 8. The full width at half maximum (FWHM) of this diffraction line was used in the Scherrer approach (equation (2)) to calculate the crystallite thickness.

Figure 8 The first order SAXS diffraction pattern of blend A upon fast and slow cooling from the melt. In order to retrieve the crystallite thickness by NMR, the spin-diffusion coefficients for the TAGs within the rigid and mobile phases have to be estimated. The spin diffusivity can be


20

Page 21 of 34

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60


expressed in terms of the local dipolar field in the case of purely dipolar spin interaction.30,44 For Gaussian and Lorentzian line shapes, the spin-diffusion coefficients can be estimated by30

=

〈 〉∆

(3)

and

= 〈 〉√ Δ

(4)

where α is the cutoff parameter of the Lorentzian line, ∆ν is the full line width at half-height, and is the mean-square distance between the nearest spins. The value of can be well represented by the weighted mean of distances between the neighboring protons of the alkyl chain, where the distance between the protons within the CH2 group and between the CH2 groups in the chain as well as the distance between the chains were taken into account30. The weighted mean square of these distances is approximately = 0.05 nm2. The spin-diffusion effect on the NMR spectra can be observed in Figure 9 where a stack of spectra as a function of diffusion time is shown. The buildup of the mobile fraction is detected with increasing diffusion time. Spectral deconvolution was carried out into three components. At very short spin-diffusion times, the NMR spectrum consists mostly of the crystalline component selected by the DQ filter. At longer times, the magnetization starts to diffuse towards the immediate proximity of the crystalline fraction, which is the interface. As time passes, the spindiffusion front reaches the end of the interfacial domains and enters the mobile phase. The interface is not necessarily a physical link between the crystalline and mobile phases, but in terms of spin transport properties it is different compared to the crystal and mobile phases. The interface protons in the first vicinity of the TAG crystallites experience a very similar molecular


21


1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Page 22 of 34

motion, which also exhibit a very similar spin transport behavior. At the other end, the protons share similar properties to the ones in the high mobility TAGs. Therefore, the interface is a highly heterogeneous medium for magnetization transport, in contrast with the “pure” phases where the fluctuations in the dipolar field and implicitly the spin-diffusivity are fairly constant. We initially considered modelling of the spin-diffusion process in the lamellar morphology using three-components. The spin-diffusion coefficient of the interface was approximated as the average of the crystalline and mobile ones. To avoid ambiguity in defining the average properties of the interface we have also considered the two-component modelling of the spin-diffusion data. In this case the interface amount obtained by spectral deconvolution was evenly distributed between the crystal and the mobile phases. The time evolution of the relative phase fractions during the spin-diffusion experiment are presented in Figure 10 for blend A under fast crystallization. Both two- and three-phase cases are presented. It can be observed that a magnetization quasi-equilibrium between the two phase fractions is reached after 100 ms for blend A (and 25 ms for blend B). The longitudinal relaxation time, T1, measured for the mobile and crystalline fractions were 1.2 s and 1.5 s, respectively (errors are less than 5 %). The spin-diffusion process, manifested in the decay and buildup curves of the rigid and mobile components, respectively, reaches quasi-equilibrium plateau during a time scale smaller than T1, so that no correction of the experimental data was considered necessary. The spin-diffusion data in Figure 10 were compared with model simulations using twoand three-component spin-diffusion models (equation (A4) in SI1 and

43

). The best modelling

results are presented in SI3 (Figure 13). We have observed that residuals of the three-component


22

Page 23 of 34

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60


fit are larger and the two-component model is a better candidate to describe the spin-diffusion data. The lamellar morphology model is justified by literature findings on crystalline TAG systems.6 For a fair comparison between NMR and SAXS results, the repeating unit domain size by NMR was considered to be the sum of the rigid, 2*interface and mobile domain thickness, which conceptually resembles the crystallite thickness defined in crystallography. The crystallite thickness obtained in this manner by NMR in the case of three-component model is 30.8 nm, and in the two-component model is 31.9 nm. We interpret this to be a relatively small difference at the crystal lengthscale, taking into account that a single TAG bilayer is of the order of 4.5 nm thick. This finding, together with the fit residuals, indicates that a two-component model can better be used to describe the spin-diffusion data in the investigated fat blends. The crystallite thickness estimated by 1H spin-diffusion NMR in the two-component model and by the Scherrer approach on SAXS data are presented in Figure 11. As a first observation, both techniques reveal significant differences between the crystallite thicknesses as a function of crystallization rate. Thinner crystallites were detected by both techniques at a higher crystallization rate, which was expected according to previous studies.24,25,45


23


1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Page 24 of 34

500.0 300.0 200.0 100.0 50.0 10.0 5.0 3.0 2.0 1.0 0.5 0.3

td [ms]

0.2 0.1 0.05 0.02

Figure 9 1H NMR stack spectra of fast crystallized blend A recorded after incremented diffusion times td in the spin-diffusion experiment using a DQ filter with τ = 9 µs and tDQ = 2 µs.


24

Page 25 of 34

1.00 0.98 0.96 0.94

normalised intensity

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60


0.92 0.90 0.88 0.86

rigid interface mobile

3 components 2 components

0.08 0.06 0.04 0.02 0.00 0

5

10

15

20

25

diffusion time1/2 [ms1/2]

Figure 10 1H spin-diffusion curves of blend A for three- and two-component models showing the decay of the intensity of the rigid (crystalline) phase and the build–up of the interface and mobile (non-crystalline) phases. (Error bars are 1 % absolute).


25


1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Page 26 of 34

Figure 11 Crystallite thickness estimated by 1H spin-diffusion NMR and by Scherrer approach on SAXS data.

CONCLUSIONS Effect of cooling rate. It is known that both cooling rate and crystallization temperature determines the type of nucleation and the polymorph formed initially.46 When cooling is fast enough to prevent direct crystallisation into a more stable polymorph, unstable forms will nucleate and grow first. For both protocols the cooling rate is sufficient to have nucleation taking place in the alpha-phase. The faster cooling with liquid nitrogen imposes a higher degree of super cooling, resulting in a higher nucleation rate, which in turn results in more, smaller


26

Page 27 of 34

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60


crystals, even upon later phase transition into the beta-prime polymorph. This was observed from the NMR and SAXS data for both fat blends presented in Figure 11. Effect of fat blend composition. Heterogeneity of fatty acids in their chain length and saturation/unsaturation introduces complexity in the TAGs behaviour in multi-component blends. Crystal packing is greatly influenced by the degree of saturation. Mono-acid saturated TAGs can pack efficiently in lamellae. Blend A is an interesterified fat blend, with tripalmitin (17%), PPL/PLP (15%) and POP/PPO (10%) as main components and a large number of other TAGs at lower concentrations account for the remaining 58%. Blend B is a hardened fat with trilaurin (15%) and LLM/LML (11%) as main components and even a larger amount (74%) of smaller concentration TAGs. Both blends have a mono-acid saturated TAG as main component, either tripalmitin or trilaurin, but in both cases, it accounts for about one sixth of the composition. In the fast cooling experiment, the fat is crash cooled far below the alpha melting point. This results in a very high nucleation rate, by which there is a very limited assorting of TAG molecules. In the following solid state phase transformation to the beta-prime, the effect on the number and thickness of crystallites remains. Only because of the slightly longer average chain length of the molecules in Blend A, the corresponding crystallites are somewhat larger than those of blend B, as can be seen in Figure 11. In the other procedure, the moderate-cooling is stopped at 20 °C, above the alpha melting point, giving a lower nucleation rate and therefore larger crystallites. In blend A the variety of TAG molecules is larger, but more important, a significant part of the blend is mono-unsaturated. This reduces the nucleation rate relative to blend B, which consists to a large extent of easily nucleating fully saturated TAGs. NMR and SAXS approaches. The small difference between the spin-diffusion NMR and the SAXS results (Figure 11) can be regarded from different perspectives. More factors can


27


1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Page 28 of 34

contribute to alter the results of both techniques. Regarding NMR, several concerns can be raised: the dimensionality of the spin-diffusion process, the accuracy of defining the spindiffusion coefficients, and the efficiency of the spin-diffusion process through possibly intermediate mobility phases (interface). It is expected that NMR reveals a smaller crystallite due to the fact that only the strongest residual dipolar couplings are edited by the DQ filter, which is related to the most precise packing of the TAGs. Even though the morphology of fat crystals is known to be lamellar, eventually the lateral size of the lamellae is finite and thus it is possible for the spin-diffusion locally to occur bi-dimensionally resulting in an overall average process dimensionality higher than 1. The possibility to encounter a semi-rigid phase especially in mixed-acid TAGs is very likely. In this case the size of the domains may slightly change and the data may require modeling considering diffusion along three mobility-different phase domains. Alternative magnetization filters28,30,34,47 could also be investigated to optimize the NMR method applicability to fat crystal networks. We will address these issues in detail in a next publication. The Scherrer method on the first order SAXS diffraction peak provides the average crystallite thickness of a crystalline TAG blend. SAXS data can be subject to extra line broadening effects caused by inner crystal stress and/or strain, which influences the calculations of the average crystallite thickness. Moreover, ambiguity exists in defining the K shape constant for fat crystals in the Scherrer formula (equation (2)). We successfully applied a 1H spin-diffusion NMR method to estimate the crystallite thickness of fat networks in TAG blends. A good agreement of results was observed between SAXS and NMR. The NMR method is of particular interest for the structure analysis of fats, especially because of the possibility to quantify the thickness of the crystalline domains in very diluted systems where scattering techniques struggle with the lack of diffracting material.


28

Page 29 of 34

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60


SUPPORTING INFORMATION Supporting information is available. SI1 Solutions of one-dimensional spin-diffusion process in lamellar morphology (finite source, finite sink). SI2 Wideline-separation NMR experiments (WISE). SI3 Results plot for the simulations of spin-diffusion curves for lamellar morphology with two and three components.

REFERENCES

(1) Walstra, P.; Kloek, W.; van Vliet, T. Fat Crystal Networks. In Crystallization processes in Fats and Lipid Systems, Marcel Dekker AG: Basel, Switzerland, 2001; pp 289-328. (2) Van Langevelde, A.; Van Malssen, K.; Driessen, R.; Goubitz, K.; Hollander, F.; Peschar, R.; Zwart, P.; Schenk, H. Structure of CnCn+2Cn-type (n = even) beta'-triacylglycerols. Acta Crystallogr. Sect. B Struct. Sci. 2000, 56 (6), 1103-1111. (3) Van Langevelde, A.; Van Malssen, K.; Hollander, F.; Peschar, R.; Schenk, H. Structure of mono-acid even-numbered beta triacylglycerols. Acta Crystallogr. Sect. B Struct. Sci. 1999, 55 (1), 114-122. (4) Kaneko, F. Polymorphism and Phase Transitions of Fatty Acids and Acylglycerols. In Crystallization Processes in Fats and Lipid Systems, Marcel Dekker AG: Basel, Switzerland, 2001; pp 53-97. (5) Kaneko, F.; Yano, J.; Sato, K. Diversity in the fatty-acid conformation and chain packing of cis-unsaturated lipids. Current Opinion in Structural Biology 1998, 8 (4), 417-425. (6) Himawan, C.; Starov, V. M.; Stapley, A. G. F. Thermodynamic and kinetic aspects of fat crystallization. Advances in Colloid and Interface Science 2006, 122 (1-3), 3-33. (7) Sato, K.; Ueno, S. Molecular Interactions and Phase Behavior of Polymorphic Fats. In Crystallization Processes in Fats and Lipid Systems, Garti, N., Sato, K., Eds.; Marcel Dekker AG: Basel, Switzerland, 2001; pp 177-209. (8) De Bruijne, D. W.; Bot, A. Fabricated Fat-based Foods. In Food Texture: Measurement and Perception, Aspen Publishers: Gaithersburg, Maryland, 1999; pp 185-227.


29


1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Page 30 of 34

(9) Bot, A.; Floter, E.; Lammers, J. G.; Pelan, E. Semi-Solid Foods. In Texture in Foods, Woodhead Publishing: Cambridge, UK., 2003; pp 350-372. (10) Marangoni, A. G.; Acevedo, N.; Maleky, F.; Co, E.; Peyronel, F.; Mazzanti, G.; Quinn, B.; Pink, D. Structure and functionality of edible fats. Soft Matter 2012, 8 (5), 1275-1300. (11) Marangoni, A. G.; Wesdorp, L. H. Structure and Properties of Fat Crystal Networks; CRC Press: 2013. (12) Acevedo, N. C.; Marangoni, A. G. Nanostructured fat crystal systems. Annu. Rev. Food Sci. Technol. 2015, 6, 71-96. (13) Peyronel, F.; Ilavsky, J.; Pink, D. A.; Marangoni, A. G. Quantification of the physical structure of fats in 20 minutes: Implications for formulation. Lipid Technol. 2014, 26 (10), 223-226. (14) Sato, K. Crystallization behaviour of fats and lipids - A review. Chemical engineering science 2001, 56 (7), 2255-2265. (15) Wright, A. J.; Narine, S. S.; Marangoni, A. G. Comparison of experimental techniques used in lipid crystallization studies. JAOCS J Am Oil Chem Soc 2000, 77 (12), 12391242. (16) Peyronel, F.; Quinn, B.; Marangoni, A. G.; Pink, D. A. Ultra small angle x-ray scattering in complex mixtures of triacylglycerols. J Phys Condens Matter 2014, 26 (46). (17) Peyronel, F.; Pink, D. A.; Marangoni, A. G. Triglyceride nanocrystal aggregation into polycrystalline colloidal networks: Ultra-small angle X-ray scattering, models and computer simulation. Curr. Opin. Colloid Interface Sci. 2014, 19 (5), 459-470. (18) Pink, D. A.; Peyronel, F.; Quinn, B.; Singh, P.; Marangoni, A. G. Condensation versus diffusion. A spatial-scale-independent theory of aggregate structures in edible oils: Applications to model systems and commercial shortenings studied via rheology and USAXS. J Phys D 2015, 48 (38). (19) Jewell, G. G.; Meara, M. L. A new and rapid method for the electron microscopic examination of fats. J Am Oil Chem Soc 1970, 47 (12), 535-538. (20) Poot, C.; Dijkshoorn, W.; Haighton, A. J.; Verburg, C. C. Laboratory separation of crystals from plastic fats using detergent solution. J Am Oil Chem Soc 1975, 52 (3), 6972. (21) Bunjes, H.; Steiniger, F.; Richter, W. Visualizing the structure of triglyceride nanoparticles in different crystal modifications. Langmuir 2007, 23 (7), 4005-4011. (22) Heertje, I.; Leunis, M. Measurement of shape and size of fat crystals by electron microscopy. LWT - Food Sci. Technol. 1997, 30 (2), 141-146.


30

Page 31 of 34

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60


(23) Acevedo, N. C.; Marangoni, A. G. Characterization of the nanoscale in triacylglycerol crystal networks. Cryst. Growth Des. 2010, 10 (8), 3327-3333. (24) Acevedo, N. C.; Marangoni, A. G. Toward nanoscale engineering of triacylglycerol crystal networks. Cryst. Growth Des. 2010, 10 (8), 3334-3339. (25) Acevedo, N. C.; Peyronel, F.; Marangoni, A. G. Nanoscale structure intercrystalline interactions in fat crystal networks. Current Opinion in Colloid & Interface Science 2011, 16 (5), 374-383. (26) Baum, J.; Pines, A. NMR studies of clustering in solids. Journal of the American Chemical Society 1986, 108 (24), 7447-7454. (27) Clauss; J; Schmidt, R.; K; Spiess; H W Determination of domain sizes in heterogeneous polymers by solid-state NMR. Acta Polymer 44, 1-17 (1993) 1993. (28) Mauri, M.; Thomann, Y.; Schneider, H.; Saalwächter, K. Spin-diffusion NMR at low field for the study of multiphase solids. Solid State Nucl Magn Reson 2008, 34 (1-2), 125141. (29) Havens, J. R.; VanderHart, D. L. Morphology of poly(ethylene terephthalate) fibers as studied by multiple-pulse 1H NMR. Macromolecules 1985, 18 (9), 1663-1676. (30) Demco, D. E.; Johansson, A.; Tegenfeldt, J. Proton spin diffusion for spatial heterogeneity and morphology investigations of polymers. Solid State Nucl Magn Reson 1995, 4 (1), 13-38. (31) Litvinov, V. M.; Xu, J.; Melian, C.; Demco, D. E.; Möller, M.; Simmelink, J. Morphology, chain dynamics, and domain sizes in highly drawn gel-spun ultrahigh molecular weight polyethylene fibers at the final stages of drawing by SAXS, WAXS, and 1H Solid-State NMR. Macromolecules 2011, 44 (23), 9254-9266. (32) Spiess, H. W. Length scales in heterogeneous polymers from solid state NMR. Macromol. Sympos. 1997, 117, 257-265. (33) Hedesiu, C.; Demco, D. E.; Kleppinger, R.; Vanden Poel, G.; Gijsbers, W.; Blumich, B.; Remerie, K.; Litvinov, V. M. Effect of temperature and annealing on the phase composition, molecular mobility, and the thickness of domains in isotactic polypropylene studied by proton solid-state NMR, SAXS, and DSC. Macromolecules 2007, 40 (11), 3977-3989. (34) Schäler, K.; Roos, M.; Micke, P.; Golitsyn, Y.; Seidlitz, A.; Thurn-Albrecht, T.; Schneider, H.; Hempel, G.; Saalwächter, K. Basic principles of static proton lowresolution spin diffusion NMR in nanophase-separated materials with mobility contrast. Solid State Nucl Magn Reson 2015, 72, 50-63. (35) Buda, A.; Demco, D. E.; Bertmer, M.; Blumich, B.; Reining, B.; Keul, H.; Hocker, H. Domain sizes in heterogeneous polymers by spin diffusion using single-quantum and


31


1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Page 32 of 34

double-quantum dipolar filters. Solid State Nuclear Magnetic Resonance 2003, 24 (1), 39-67. (36) Voda, M. A.; Demco, D. E.; Voda, A.; Schauber, T.; Adler, M.; Dabisch, T.; Adams, A.; Baias, M.; Blümich, B. Morphology of thermoplastic polyurethanes by1H spin-diffusion NMR. Macromolecules 2006, 39 (14), 4802-4810. (37) Assink, R. A. Nuclear spin diffusion Macromolecules 1978, 11 (6), 1233-1237.

between

polyurethane

microphases.

(38) Meyer, H. W.; Schneider, H.; Saalwächter, K. Proton NMR spin-diffusion studies of PSPB block copolymers at low field: Two- vs three-phase model and recalibration of spindiffusion coefficients. Polym J 2012, 44 (8), 748-756. (39) Herrera, M. L.; Hartel, R. W. Effect of processing conditions on crystallization kinetics of a milk fat model system. JAOCS J Am Oil Chem Soc 2000, 77 (11), 1177-1187. (40) Campos, R.; Narine, S. S.; Marangoni, A. G. Effect of cooling rate on the structure and mechanical properties of milk fat and lard. Food Research International 2002, 35 (10), 971-981. (41) Voda, M. A.; Goudappel, G. J.; den Adel, R.; van Duynhoven, J. Domain Sizes in Triglyceride Crystal Networks by 1H Spin-diffusion NMR. In Magnetic Resonance in Food Science: Food for Thought, The Royal Society of Chemistry: 2013; pp 61-71. (42) Wang, J. On the determination of domain sizes in polymers by spin diffusion. J Chem Phys 1996, 104 (12), 4850-4858. (43) Buda, A.; Demco, D. E.; Bertmer, M.; Blumich, B.; Litvinov, V. M.; Penning, J. P. General analytical description of spin-diffusion for a three-domain morphology. Application to melt-spun nylon 6 fibers. Journal of Physical Chemistry B 2003, 107 (22), 5357-5370. (44) Cheung, T. T. P. Spin diffusion in NMR in solids. Physical review. B, Condensed matter 1981, 23 (3), 1404-1418. (45) Maleky, F.; Acevedo, N. C.; Marangoni, A. G. Cooling rate and dilution affect the nanostructure and microstructure differently in model fats. Eur. J. Lipid Sci. Technol. 2012, 114 (7), 748-759. (46) Van Malssen, K.; Peschar, R.; Schenk, H. Real-time X-Ray powder diffraction investigations on cocoa butter. I. Temperature-dependent crystallization behavior. JAOCS J Am Oil Chem Soc 1996, 73 (10), 1209-1215. (47) Saalwächter, K. Proton multiple-quantum NMR for the study of chain dynamics and structural constraints in polymeric soft materials. Prog. Nucl. Magn. Reson. Spectrosc. 2007, 51 (1), 1-35.


32

Page 33 of 34


1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 ACS Paragon Plus Environment

33


1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Page 34 of 34

For Table of Contents Use Only

Quantitative assessment of triacylglycerol crystallite thickness by 1H spin-diffusion NMR Adrian Voda, Ruud den Adel, Kees van Malssen, John van Duynhoven

Synopsis: We demonstrated the potential use of 1H spin-diffusion NMR to estimate the thickness of fat crystallites. A good agreement of results was observed between SAXS and NMR. The NMR method is of particular interest for the structure analysis of fats, especially in diluted systems where scattering techniques struggle with the lack of diffracting material.


34

Quantitative Assessment of Triacylglycerol Crystallite Thickness by 1H

Recommend Documents