Raman Spectroscopy As a Non-Destructive Tool to Quantify The

Analytical Methods Accepted Manuscript

This is an Accepted Manuscript, which has been through the Royal Society of Chemistry peer review process and has been accepted for publication.

Accepted Manuscripts are published online shortly after acceptance, before technical editing, formatting and proof reading. Using this free service, authors can make their results available to the community, in citable form, before we publish the edited article. We will replace this Accepted Manuscript with the edited and formatted Advance Article as soon as it is available.

You can find more information about Accepted Manuscripts in the Information for Authors.

Please note that technical editing may introduce minor changes to the text and/or graphics, which may alter content. The journal’s standard Terms & Conditions and the Ethical guidelines still apply. In no event shall the Royal Society of Chemistry be held responsible for any errors or omissions in this Accepted Manuscript or any consequences arising from the use of any information it contains.

www.rsc.org/methods Page 1 of 19 Analytical Methods

1 2 3 4 Raman Spectroscopy as a non-destructive tool to quantify the 5 6 comonomer content in ethylene/1-olefin copolymers 7 8 9 * 10 Abhishek Sanoria, Tobias Schuster, Robert Brüll 11 12 Fraunhofer Institute for Structural Durability and System Reliability, Division 13 Plastics, Group Material Analytics, Schlossgartenstrasse 6, 64289 Darmstadt, Germany 14 15 *Corresponding author: Dr. Robert Brüll, [email protected] 16 17 18 19 Abstract: 20 Manuscript 21 Ethylene/1-olefin copolymers, ranging from 1-butene to 1-octadecene, have been studied with Raman 22 23 spectroscopy and the spectral changes upon comonomer incorporation have been analyzed. The 24 25 Raman spectra of ethylene/1-olefin copolymers are characterized by a band ensemble in the lower 26 27 frequency region below 1000 cm-1 which is a characteristic scattering from the comonomer segments. 28 29 30 The spectra show bands due to scattering from the polyethylene sequences which also reflect the effect 31 32 of the morphological changes occurring with the increasing comonomer content. By correlating the 33 34 bands characteristic of the comonomer to those of the ethylene sequences, an internal intensity Accepted 35 36 standard was identified. This also caters to the effect of the experimental parameters such as sample 37 38 focus and spectral acquisition time for calibration of the comonomer content. This approach was also 39 40 extended to amorphous ethylene/norbornene copolymers and a similar method for quantification was 41 42 43 developed. The major advantages of this method are the minimum amount of sample preparation, and 44 45 the low sample size, enabling even micrograms of the sample to be measured. Due to these, 46 Methods 47 quantification with Raman spectroscopy is uniquely suited for online quality control and liquid 48 49 chromatography which require a fast and robust analytical methodology. 50 51 52 53 54 Keywords: 55 56 57 Raman spectroscopy, polyethylene, ethylene/1-olefins, comonomer content 58 59 Analytical 60

Analytical Methods Page 2 of 19

1 2 3 4 Introduction: 5 6 7 8 Polyethylene (PE) is one of the most important industrial polymers finding a wide range of 9 10 applications in durable and consumable goods, with a global consumption soaring to 76 million tons in 11 12 13 2010. An excellent cost/performance ratio, recyclability, ease of synthesis and processability have

14 1, 2 15 made the market for these polymers to grow continuously . PE can be classified into the 16 17 homopolymer and semicrystalline copolymers of ethylene with 1-olefins containing up to 20 mol % of 18 19 comonomer (Linear Low Density Polyethylene, LLDPE). Typically, 1-butene, 1-hexene and 1-octene 20 Manuscript 21 are used as the comonomer with ethylene. Until the late 1980’s PE was synthesized by free radical 22 23 polymerization or by heterogeneous transition metal catalysts. The early 1990’s saw the addition of 24 25 1-5 26 metallocene catalysts to the portfolio of catalyst systems which allow the production of copolymers

27 6 28 which are narrowly distributed with regard to molar mass and comonomer incorporation . Such, and 29 30 further advances in the catalyst and process technology thus helped in gaining a better control over the 31 32 structure dependent performance attributes of PE copolymerized with 1-olefins such as toughness, 33 Accepted 34 crack resistance and optical properties 6. The comonomer content is a fundamental molecular 35 36 parameter, which strongly influences the macroscopic polymer properties. As a consequence, there 37 38 39 continues to be a need for analytical techniques, which are fast and require minimum effort with 40 41 regard to sample preparation for quantification. 42 43 44 45 Nuclear Magnetic Resonance (NMR) has been widely used for structure elucidation of ethylene/1- 46 Methods 47 olefin-copolymers and to quantify the comonomer content 7. NMR is specific in terms of acquiring 48 49 information about the microstructure such as tacticity 8, inverse insertion 9 and comonomer sequence 50 51 10 52 distribution . Being an absolute technique, NMR does not require a calibration, however quantitative 53 54 measurements require significant amounts of sample, with the mass range depending on 55 56 instrumentation and experimental parameters. Additionally, sample preparation may pose a bottleneck 57 58 in high throughput environments. Fourier Transform Infrared Spectroscopy (FT-IR) has also been 59 Analytical 60 shown capable of such quantification but has several limitations 11, 12. The measurement can be carried

out in Attenuated Total Reflectance (ATR) and transmission mode. The first one is a surface 2

Page 3 of 19 Analytical Methods

1 2 3 technique, as the spectral information is retrieved from a few microns below the sample surface. The 4 5 obtained spectra are affected by a number of experimental parameters such as the applied pressure, the 6 7 13 8 surface area of contact and the change in the refractive index with the penetration into the sample . 9 10 However, quantitative information can only be retrieved from measurements in transmission mode 11 12 which requires significant quantities of sample, typically a few grams, and intensive sample 13 14 preparation. 15 16 17 18 Raman spectroscopy is another method of vibrational spectroscopy, which is sensitive to composition 19 20 Manuscript 21 and morphological parameters like orientation and degree of crystallinity. The short spectral 22 23 acquisition times and minimal sample preparation make it a predestined technique to analyze polymer 24 25 blends, copolymers and polymer composites. The interaction of PE with highly conductive fillers such 26 27 as carbon nanotubes has been studied by analyzing the band shifts in the Raman spectra of the 28 29 nanotubes upon intercalation of PE in the nanotube bundles 14, 15. PE/ polypropylene (PP) blends have 30 31 also been quantified and being thermodynamically immiscible 16, the phase composition and 32 33 17 34 crystallinity for the individual blend components has also been determined . Accepted 35 36 37 The Raman spectrum of PE is well understood and the fundamental modes of vibration have been 38 18-20 39 assigned . The internal mode region of the spectrum has been grouped into three frequency 40 −1 −1 41 regions, namely the C-C stretching region between 1000 cm and 1200 cm , the CH2 twisting modes 42 −1 −1 −1 21 43 near 1300 cm and the CH2 bending modes between 1400 cm - 1500 cm . Subtle changes in the 44 45

Raman spectra of olefin copolymers upon comonomer incorporation (1-olefins) have been observed, Methods 46 47 but have not been developed into a tool to quantify the comonomer content. For copolymers of 48

49 -1 -1 50 propylene with 1- olefins, changes in the intensity of the bands at 809 cm and 841 cm were 51 22 52 observed with increasing content of the incorporated comonomer . The first of these bands 53 54 corresponds to vibrations of the helically oriented PP chains in the crystalline phase, while the second 55 56 one is associated with vibrations of the helically aligned PP chains with significant amounts of 57 58 conformational defects, localized in the amorphous phase. When small amounts (0-20 mol %) of 1- 59 Analytical 60 olefinic comonomers are copolymerized with ethylene, the changes reflected in the Raman spectra are

minute, and these finite details have not gained much attention to be useful for any quantification. For 3

Analytical Methods Page 4 of 19

1 2 3 the case of ethylene/1-hexene copolymers changes in the Raman spectra were identified, but have not 4 5 been developed as a tool for quantification of the comonomer content 23. 6 7 8 9 10 So far, analysis of PE by Raman spectroscopy has been limited to deriving the density, degree of 11 12 crystallinity, and melting point from multivariate analysis of the Raman spectra 24, 25. Mauler et al. 13 14 investigated the influence of comonomer content on the degree of crystallinity and dynamic 15 16 mechanical properties of ethylene/1-octene copolymers 26. The Raman spectra were analyzed and a 17 18 decrease in the degree of crystallinity was observed. Henceforth, developing a fast, accurate and robust 19 20 Manuscript 21 method to quantify the comonomer content in ethylene copolymers by Raman spectroscopy would 22 23 extend this technique further than crystallinity and density measurements, making it suitable for a 24 25 rapid identification of minute sample amounts. The need for such a tool exists for quality control in 26 27 polymer synthesis. A further area where a compositional analysis of small amounts of sample is 28 29 required can be found in liquid chromatography, where, after removing the chromatographic solvent, 30 31 trace amounts of the remaining polymer have to be analyzed. The non-destructive nature of Raman 32 33 34 spectroscopy renders it also a good potential in forensic analysis and the analysis of very small Accepted 35 36 inhomogenities in processed samples e.g., multi-layer films. In this study, we exploit the subtle 37 38 changes in the Raman spectra of PE upon comonomer incorporation and understand these changes to 39 40 present a new method establishing Raman spectroscopy as a tool for comonomer quantification for 41 42 several industrially important ethylene/1-olefins. 43 44 45 46 Methods 47 Experimental: 48 49 Samples: 50 51 52 Samples used are detailed as in Table I below. All samples were synthesized using metallocene 53 54 catalyst systems. 55 56 57 58 59 Analytical 60

Page 5 of 19 Analytical Methods

1 2 3 Table I: Sample description 4 5 6 Copolymer type Notation Content [Mol % of Dispersity Supplier 7 1-olefin 8 comonomer 9 determined 10 through NMR] 11 E 1.57 – 10.32 Braskem, Brazil 12 ethylene /1-butene BUT 3.2-4.3 1.8-2.9 O. Boyron, Univ. of 13 ethylene /1-hexene EHEX 2.2 – 20.7 14 Lyon, Lyon France EOCT 2-2.2 The Dow Chemical 15 ethylene /1-octene 0 – 13.9 16 Company, USA 17 ethylene /1- 1.9-2.6 O. Boyron, Univ. of E 1.7-3.3 18 octadecene OCTD Lyon, Lyon France 19 1.8-1.9 Topas Advanced 20 ethylene

E 24.5 – 60.75 Polymers GmbH, Manuscript /norbornene NOR 21 Germany 22 23 24 25 26 27 Raman Spectroscopy: 28 29 Raman measurements were carried out using a WITec alpha 500 spectroscopy system. For 30 31 measurements on pellets/powders, a 20× (NA 0.4) microscope objective was used to focus the laser 32 33 beam and to collect the scattered light. Laser radiation having a wavelength of 532 nm with a power of Accepted 34

35 −1 36 about 50 mW was employed as an excitation source. The scattered light was analyzed in a 600 mm 37 −1 38 grating spectrometer at a spectral resolution of about 1 cm . After focusing the sample manually with 39 40 the microscope, the laser spot was positioned to accumulate a single spectrum with an acquisition time 41 42 of 2 s and 10 accumulations were carried out per measurement. Each spectrum was smoothened by 43 44 using a Savitzky-Golay method and background subtracted. 45 46 Methods 47 48 Results and Discussions: 49 50 The Raman spectra of EOCT copolymers with a comonomer content ranging from 0 – 13.9 mol % in the 51

52 range of the C-C stretching and the deformation vibrations of CH2 and CH3 groups are shown in Fig. 53 54 1. 55 56 57 58 59 Analytical 60

Analytical Methods Page 6 of 19

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 Manuscript 21 22 23 24 25 26 27 28 29 30 31 32 33 34 Accepted 35 36 37 38 39 40 41 42 Fig. 1: a) Raman spectra of E with varying 1-octene content and b) Raman spectra of E showing the band 43 OCT OCT 44 ensemble below 1000 cm-1 45 46 Methods 47 The spectra show pronounced differences with increasing content of the comonomer. The 48 49 characteristic bands for the ethylene sequences show a marked decrease in intensity and 50 51 simultaneously, the intensity of the band at 1416 cm−1 which has been widely used to measure the 52 53 degree of crystallinity 21 also decreases. The bands in the C-C stretching region assigned at 1058 cm−1, 54 55 −1 −1 −1 −1 20 56 1126 cm and 1166 cm , the twisting region at 1295 cm and the bending region at 1439 cm 57 −1 58 follow a similar trend. Interestingly, the frequency region below 1000 cm shows a band ensemble 59 Analytical 60 which is characteristic for the 1-octene comonomer as can be recognized from the spectrum of poly-1-

Page 7 of 19 Analytical Methods

1 2 3 octene. This band ensemble, normalized with respect to the highest intensity band at 1058 cm−1, shows 4 5 a significant increase in intensity with comonomer incorporation (Fig. 1 b). 6 7 8 9 10 A method to predict the crystallinity of PE by Raman spectroscopy was developed by investigating the 11 12 spectra of bulk-crystallized PE 19. The spectra were analyzed to determine the content of the three 13 14 distinct polymorphs, namely the orthorhombic crystalline phase, a melt like amorphous phase, and a 15 16 disordered anisotropic phase. The intensity of the band at 1416 cm-1 was used to deduce the relative 17 18 fraction of segments present in the orthorhombic phase (α ). The relative amount of segments forming 19 c 20 Manuscript 21 the liquid like amorphous phase (αa) was obtained from two bands centered in the twisting region at 22 -1 -1 23 1303 cm or in the stretching region at 1080 cm . The interfacial phase (αb) was proposed as a domain 24 25 wherein the chains are in the extended trans-configuration, but have lost their lateral order, and the 26

27 relative amount of segments forming this phase was calculated as αb = 1- αa- αc. With increasing 1- 28 29 octene content the Raman band characteristic for measuring the crystallinity (αc) in the PE segments at 30 31 1416 cm−1 decreases in intensity and remains as a small shoulder in the sample with 8.5 mol % 1- 32 33 −1 34 octene. Similarly, with the increase in comonomer content the band at 1166 cm merges with the Accepted 35 −1 36 more intense band at 1126 cm . The alkyl branches of the comonomer are not able to enter the crystal 37 22 38 lattice , and hence, an increase in the fraction of αa and αb is observed. 39 40 41 42 Thus, the Raman spectra reflect the effect of the decrease in band intensity of the characteristic bands 43 44 from the PE segments due to the decrease in the ethylene content, the effect of the comonomer 45 46 Methods 47 segments seen as a band ensemble and the associated changes in the phase composition due to 48 49 comonomer incorporation. Broadening has also been observed in these characteristic PE bands, which 50 51 is deduced to be due to the increasing content of 1-octene which shows several bands overlapping in 52 53 the same frequency region as the PE segments (Fig. 1b). The above three effects limit the applicability 54 55 of using solely the Raman bands of the PE segments to determine the comonomer content. However, 56 57 before selecting suitable bands for quantification it has to be taken into account that the intensity of the 58 59 Analytical 60 Raman spectra varies to a great extent with experimental parameters, taking in the first instance the

Analytical Methods Page 8 of 19

1 2 3 spectral acquisition time and the sample focus/depth 27.Fig. 2 shows the Raman spectra of linear PE 4 5 measured with different acquisition times after being focused with the microscope. 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 Manuscript 21 22 23 24 Fig. 2: Variation in signal intensity of the Raman spectra of linear PE with spectral acquisition times. 25 26 27 28 29 A clear dependence of the signal intensity on the acquisition time can be seen, and the intensity of the 30 31 Raman signal increases with the spectral acquisition time. Additionally, in order to maximize the 32 33

signal intensity and reduce the signal/noise ratio, spectral acquisition has to be preceded by proper Accepted 34 35 focusing of the sample. The latter is, however, a very subtle parameter often depending on the skill of 36 37 38 the operator, and even a few microns shift in the sample focus can have a significant influence on the 39 40 signal intensity. This in turn would have a direct impact on the height/area of individual bands, thus 41 42 again reducing their applicability as the sole criterion to determine the comonomer content in 43 44 samples.Fig. 3 shows this effect where a sample was focused manually (same spectral acquisition 45 Methods 46 time) and then Raman spectra were acquired at one micrometer levels above and below the sample. 47 48 The intensity of the bands at 1295 and 1439 cm-1 have been plotted with the variation in sample 49 50 51 height. 52 53 54 55 56 57 58 59 Analytical 60

Page 9 of 19 Analytical Methods

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 Fig. 3: Effect of sample focusing on the scattering intensity the sample 17 18 19 A significant variation in band intensity is seen with the shift in focus above and below the sample. 20 Manuscript 21 22 Using the spectral acquisition time of 2 seconds and focusing the samples manually, the intensity of 23 24 -1 -1 the band at 885 cm in the band ensemble of EOCT below 1000 cm was correlated with the 25 26 comonomer content as shown inFig. 4 below. 27 28 29 30 31 32 33 34 Accepted 35 36 37 38 39 40 41 42 43 Fig. 4: Band intensity I as a function of 1-octene content 44 855 45 Methods 46 The effect of the experimental parameters on the intensity of the band at 885 cm-1 for determining the 47 48 comonomer content is evident inFig. 4 but the correlation with the comonomer content is non-linear. 49 50 51 Thus, for developing a method to correlate these spectral changes upon comonomer incorporation and 52 53 also eliminating effect of the experimental parameters on the spectra, an internal intensity 54 55 standard/reference for normalizing the intensity of the band ensemble is required. Hence, the ratio of 56 57 the band intensities of the characteristic bands from the ethylene sequences to the characteristic bands 58 59 emerging from the comonomer was investigated to find an internal standard which also minimizes the Analytical 60 effect of the spectral overlap due to scattering from the comonomer sequences. Normalizing the band

Analytical Methods Page 10 of 19

1 2 3 intensity ratio would cancel out the effect of sample focusing, and thus, reduce the number of 4 5 parameters to find a suitable correlation for quantification. The high intensity bands characteristic of 6 7 −1 -1 −1 8 the PE segments at 1058 cm (C-C stretching), 1126 cm (C-C stretching), 1295 cm (C-C twisting) 9 −1 10 and at 1439 cm (C-C bending) were selected. The band ensemble characteristic for the comonomer 11 12 from 700 cm−1- 950 cm−1 shows four major bands at 844 cm−1, 855 cm−1, 875 cm−1 and 885 cm−1 13 14 respectively. The ratio of the intensity of the strongest band amongst these at 885 cm−1 to these 15 16 characteristic PE bands was probed to determine a correlation which accounts for the effect of 17 18 increasing comonomer content, the accompanying decrease in the ethylene content and intrinsic 19 20 Manuscript 21 changes in the ratio of polymorphs. The band ratios with the respective standard deviation in the 22 23 measurements in correlation with the comonomer content are presented inFig. 5. 24 25 26 27 28 29 30 31 32 33 34 Accepted 35 36 37 38 39 40 41 42 43 44 45 Methods 46 47 Fig. 5: Band intensity ratios a) I855/ I1058 b) I855/I1126 c) I855/I1295 d) I855/I1439 plotted as a function of 1-octene content 48 49

50 −1 51 As seen from Fig. 5, the band at 1295 cm in the C-C twisting region yielded the best correlation with 52 2 −1 53 an R value of 0.9949. The band at 1439 cm in the C-C bending region shows broadening and 54 55 overlaps with a characteristic band of 1-octene at 1444 cm−1 yielding an R2 value of 0.9791. This band 56 57 is also seen to merge with the one at 1416 cm−1 used to determine the crystallinity of PE as discussed 58 59 previously, and the combined effect of these two bands leads to a non-linear correlation. The band at Analytical 60 1058 cm−1 in the C-C stretching region yields an R2 value of 0.9841, but starts to develop a shoulder

Page 11 of 19 Analytical Methods

1 2 3 due to merging with the characteristic bands of 1-octene at 1059 and 1080 cm−1 when the 1-octene 4 5 content is increased. The other band in the stretching region at 1126 cm-1 gives an R2 value of 0.9878 6 7 -1 8 also merging with the band characteristic for the 1-octene comonomer at 1140 cm showing non- 9 10 linearity. Most of the characteristic Raman bands for 1-octene are in close vicinity to those for PE, 11 12 however, the effect on the Raman scattering intensity of the C-C twisting mode was the least and was 13 14 observed to fit excellently correlating with the comonomer content. 15 16 17 18 To verify the reproducibility and the robustness of this method, ten independent measurements were 19 20 Manuscript 21 carried out per sample and a collective average of the band ratios was plotted. The deviation from the 22 23 linear fitting was observed to be negligible, verifying the applicability and robustness of the method. 24 25 -1 26 The Raman spectra of EBUT with varying content of 1-butene in the frequency range 600-1070 cm are 27 28 shown in Fig. 6. 29 30 31 32 33 34 Accepted 35 36 37 38 39 40 41 42 43 44 45 46 Methods 47 Fig . 6: Raman spectra of E normalized to the high intensity band at 1058 cm-1 48 BUT 49 50 51 With increasing comonomer content, similar changes in the Raman spectra as observed for E can 52 OCT 53 54 be seen for EBUT with the decrease in intensity of ethylene sequences and the appearance of a band 55 -1 56 ensemble in the region below 1000 cm . The latter shows a clear increase in intensity with increasing 57 58 comonomer content, but was difficult to resolve in terms of individual bands due to the presence of 59 Analytical 60 several bands having similar intensity. On a closer inspection, the band at 869 cm-1 could be identified

in the ensemble showing the highest variation in intensity with comonomer content. Using an 11

Analytical Methods Page 12 of 19

1 2 3 analogous approach as for EOCT, the bands from the three different modes of vibration for the ethylene 4 5 sequences were selected to develop a method to quantify the 1-butene comonomer. The intensity of 6 7 -1 8 the band at 869 cm was chosen for drawing out a correlation with the bands from the ethylene 9 10 sequences to deduce a method for quantifying the content of EBUT.Fig. 6 shows the intensity ratio of 11 12 the band at 869 cm-1 to those arising from the ethylene. 13 14 15 16 17 18 19 20 Manuscript 21 22 23 24 25 26 27 28 29 30 31 32 33 34 Accepted 35 36 37 Fig. 7: Band intensity ratios a) I869/ I1058 b) I869/I1126 c) I869/I1295 d) I869/I1439 as a function of 1-butene content 38 39 40 41 In this case as well, the intensity ratio of the band at 869 cm-1 to the band in the C-C twisting region at 42 43 1295 cm-1 showed a more accurate correlation with the 1-butene content than the bands in the 44 45 −1

stretching and bending regions. The band at 1439 cm in the C-C bending region was especially Methods 46 47 2 48 influenced by scattering from the 1-butene sequences giving an R value of 0.9696. Hence, the effect 49 50 of the spectral overlap of the scattering from the 1-butene sequences with the ethylene sequences and 51 52 its relation with the characteristic band of the 1-butene sequences found the best synchronization with 53 -1 2 54 the band at 1295 cm yielding an R value of 0.9947 for quantification of EBUT. 55 56 57 58 The Raman spectra of E with varying content of 1-hexene in the frequency range from 600-1070 59 HEX Analytical 60 cm-1 are shown inFig. 8.

Page 13 of 19 Analytical Methods

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 Manuscript 21 Fig. 8: Raman spectra of E normalized to the high intensity band at 1058 cm-1 22 HEX 23 24 25 E , similar to E and E also shows a band ensemble in the low frequency region below 1000 26 HEX OCT BUT

27 -1 28 cm which grows in intensity with comonomer incorporation. The ensemble shows the presence of 29 -1 -1 -1 30 two prominent bands at 879 cm and 893 cm , with the one at 893 cm being dominant, and in 31 32 addition showing a variation in intensity with 1-hexene content. Therefore, the latter was selected as 33 Accepted 34 representative of 1-hexene and was used to find a correlation with the characteristic bands in the 35 36 different vibrational regions for the ethylene sequences as selected earlier. The effect of band overlap 37 38 and changes in the bands from the ethylene sequences was studied and the best fit for obtaining the 1- 39 40 41 hexene content through the Raman spectra of EHEX was identified (Fig. 9). 42 43 44 45 46 Methods 47 48 49 50 51 52 53 54 55 56 57 58 59 Analytical 60

Analytical Methods Page 14 of 19

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 Manuscript 21 22 23 24 25 26 27 28 Fig. 9: Band intensity ratios a) I893/ I1058 b) I893/I1126 c) I893/I1295 d) I893/I1439 as a function of 1-hexene content 29 30 In this case, similar to EOCT, the best correlation was found with the intensity ratio of the band at 893 31 32 cm-1 characteristic of the 1-hexene comonomer to that of the band in the C-C twisting region at 1295 33 Accepted 34 cm-1 yielding an R2 value of 0.9922. 35 36 37 The Raman spectra of E with varying content of 1-octadecene in the frequency range from 600- 38 OCTD 39 -1 40 1070 cm are shown in Fig. 10. 41 42 43 44 45 46 Methods 47 48 49 50 51 52 53 54 55 56 57 58 59 Analytical 60

Page 15 of 19 Analytical Methods

1 2

3 -1 Fig. 10: Raman spectra of E normalized with regard to the high intensity band at 1058 cm 4 OCTD 5 6 7 The spectra show the appearance of a band ensemble in the low frequency region below 1000 cm-1 as 8 9 seen previously for other 1-olefins. The increasing content of 1-octadecene is also characterized by the 10 11 12 increase in the intensity of this ensemble which becomes increasingly prominent with the increasing 13 14 comonomer content. Quantification of EOCTD was carried out using a similar approach of choosing the 15 16 band ratios of the most prominent band in the ensemble to the characteristic bands from the scattering 17 18 through the ethylene sequences. The band ensemble in this case was clearly seen, and the highest 19 20 -1

intensity band at 896 cm was used as a representative of the scattering from 1-octadecene and its Manuscript 21 22 correlation with the characteristic bands from the ethylene sequences was determined (Fig.11). 23 24 25 26 27 28 29 30 31 32 33 34 Accepted 35 36 37 38 39 40 41 42 43 44 45

Methods 46

47 Fig.11: Band intensity ratios a) I896/ I1058 b) I896/I1126 c) I896/I1295 d) I896/I1439 as a function of 1-octadecene content 48 49 50

51 In this case for EOCTD as well, the best linear correlation for determining the comonomer content was 52 53 observed with the band in the C-C twisting region at 1295 cm-1giving an R2 value of 0.992. Here too, 54 55 56 the effect of the spectral overlap of the scattering through 1-octadecene is pronounced in the C-C 57 58 bending region. 59 -1 Analytical 60 Hence, it can be concluded that the band in the C-C twisting region at 1295 cm can be used as an

internal intensity standard for normalizing the intensity of the band ensemble to quantify the

Analytical Methods Page 16 of 19

1 2 3 comonomer content of LLDPE. The Raman spectra of the comonomers from 1- butene to 1- 4 5 octadecene have been studied and the intensity of this band to be used as an internal standard 6 7 8 appropriate for quantification has been verified. 9 10 11 12 Quantification of Cyclic Olefin Copolymer (COC) norbornene 13 14 15 16 An interesting question is, if such a method to quantify the comonomer content can also be applied to 17 18 ethylene copolymers of low crystallinity. Cyclic Olefin Copolymers (COCs) are commercially 19 20 Manuscript 21 important materials, which due to their superior optical, mechanical, thermal and water barrier 22 23 properties find a wide range of applications ranging from packaging films, lenses, vials, displays, and 24 25 medical devices. Norbornene is frequently used as a comonomer with ethylene, including others such 26 27 as cyclopentene and dicyclopentadiene 28, 29. Copolymers of ethylene with straight chain aliphatic 1- 28 29 olefins generally have lower comonomer content (up to 20 mol %) compared to COC’s for which the 30 31 commercially relevant comonomer content ranges from 20-60 mol %. Due to the high comonomer 32 33 34 content, these materials are fully amorphous, and extending the applicability of this method of Accepted 35 36 quantification to such amorphous materials would add another line of commercially important 37 38 materials quantifiable nondestructively. At such a concentration of the comonomer, the spectrum of 39 40 the comonomer starts dominating, suppressing and overlapping with the dominant bands observed due 41 42 to the ethylene sequences (Fig. 10). 43 44 45 46 Methods 47 48 49 50 51 52 53 54 55 56 57 58 59 Analytical 60

Fig. 12: Raman spectra of ENOR 16

Page 17 of 19 Analytical Methods

1 2 3 4 5 The bridged cyclic structure in the copolymer gives rise to several high intensity bands from 500 cm-1- 6 7 -1 8 1500 cm . The bands solely characteristic of the ethylene sequences are not segregable easily apart 9 -1 10 from the band at 1439 cm which can still be seen prominently. To determine the effect of the 11 12 decrease in the percentage of the ethylene sequences and the increase in the norbornene content 13 14 through the Raman spectra, the intensity of the band at 1439 cm-1 was selected and plotted with 15 16 respect to the intensity of the band at 1313 cm-1 attributed to the cyclic comonomer. 17 18 19 20 Manuscript 21 22 23 24 25 26 27 28 29 30 31 32 33 34 Accepted 35 36 Fig. 13: Band intensity ratio I / I as a function of the norbornene content 37 1439 1313 38 39 A good linear correlation (Fig. 13) substantiates the fact that the two bands chosen are appropriate to 40 41 quantify the content of norbornene in COCs. 42 43 44 45 46 Conclusions: Methods 47 48 Quantification of the comonomer content in linear low density polyethylene is of enormous 49 50 51 importance for many industrial applications such as quality control and sample identification. NMR 52 53 spectroscopy has been widely used for structure elucidation of copolymers, but quantitative 54 55 measurements require significant amount of sample and sample preparation may pose a bottleneck, 56 57 limiting its scope for applications in high throughput analysis and in cases, where measurements are 58 59 limited by the amount of sample available. In this study, the potential of Raman spectroscopy has been Analytical 60 exploited as a fast and nondestructive alternative for comonomer quantification requiring no sample

Analytical Methods Page 18 of 19

1 2 3 preparation. The Raman spectra of ethylene/1-olefin copolymers, ranging from 1-butene to 1- 4 5 octadecene show a band ensemble in the lower frequency region below 1000 cm-1 which results from 6 7 8 the scattering from the comonomer sequences and increases in intensity with increasing comonomer 9 10 content. Upon comonomer incorporation the spectra reflect the combined effect of the decrease in 11 12 ethylene content, associated changes in the phase composition of polyethylene and increased scattering 13 14 from the comonomer sequences. The band ensemble characteristic for the comonomer itself cannot be 15 16 used for quantification since the intensity of the spectra is affected by the sample focus and the 17 18 spectral acquisition time. These effects, together with the spectral changes upon comonomer 19 20 Manuscript 21 incorporation, have been analyzed to identify an internal standard for normalizing the intensity of the 22 -1 23 band ensemble. The intensity of the band at 1295 cm in the C-C twisting region was found to be 24 25 appropriate for this purpose also accounting for the effect of the experimental parameters in the 26 27 spectra. This method has also been extended for determining the comonomer content of cyclic olefin 28 29 copolymers having low crystallinity where the bands arising due to scattering from the comonomer 30 31 sequences start dominating. Hence, in this study the Raman spectra of copolymers of ethylene with 1- 32 33 34 olefins ranging from 1-butene to 1-octadecene and the cyclic norbornene have been analyzed and a Accepted 35 36 method has been developed to quantify the comonomer content. Due to the minimal sample 37 38 preparation, fast spectral acquisition time and its nondestructive nature, the Raman spectroscopic 39 40 approach would be very useful for quality controls. A strong need for such a technique also exists in 41 42 liquid chromatography where the amounts of sample acquired are regularly very small. 43 44 45 46 Methods 47 Acknowledgements: 48 49 The author would like to acknowledge the support of Dr. Tibor Macko, Dr. Dibyaranjan Mekap and 50 51 Dr. Frank Malz for providing the samples and for valuable discussions for the study. 52 53 54 55 56 References 57 58 1. W. Kaminsky, Macromolecular Chemistry and Physics, 2008, 209, 459-466. 59 Analytical 60 2. P. S. Chum and K. W. Swogger, Progress in Polymer Science, 2008, 33, 797-819. 3. W. Kaminsky, Journal of Polymer Science Part A: Polymer Chemistry, 2004, 42, 3911-3921. 4. J. Coates, in Encyclopedia of Analytical Chemistry, John Wiley & Sons, Ltd, Editon edn., 2000.

Page 19 of 19 Analytical Methods

1 2 3 5. J.-S. Wang and K. Matyjaszewski, J Am Chem Soc, 1995, 117, 5614-5615. 4 6. P. Chum, W. Kruper and M. Guest, Advanced Materials, 2000, 12, 1759-1767. 5 6 7. G. R. Hatfield, W. E. Killinger and R. C. Zeigler, Analytical Chemistry, 1995, 67, 3082-3085. 7 8. V. Busico and R. Cipullo, Progress in Polymer Science, 2001, 26, 443-533. 8 9. Z. Zhou, R. Cong, Y. He, M. Paradkar, M. Demirors, M. Cheatham and A. W. deGroot, 9 Macromolecular Symposia, 2012. 10 10. J. C. Randall, Journal of Macromolecular Science—Reviews in Macromolecular Chemistry and 11 Physics, 1989, 29, 201-317. 12 13 11. J. P. Blitz and D. C. McFaddin, J Appl Polym Sci, 1994, 51, 13-20. 14 12. M. C. Harvey and A. D. Ketley, J Appl Polym Sci, 1961, 5, 247-250. 15 13. Z. M. Khoshhesab, 2012. 16 14. T. McNally, P. Pötschke, P. Halley, M. Murphy, D. Martin, S. E. Bell, G. P. Brennan, D. Bein, P. 17 Lemoine and J. P. Quinn, Polymer, 2005, 46, 8222-8232. 18 15. L. Wei, W. Jiang, K. Goh and Y. Chen, Polymer Engineering & Science, 2014, 54, 336-344. 19 16. N. Livanova, E. Popova, O. Ledneva and A. Popov, International Journal of Polymeric 20 Manuscript 21 Materials, 2000, 47, 279-291. 22 17. E. A. S. Yu.V. Zavgorodnev, G.Yu. Nikolaeva, K.A. Prokhorov,, L. A. N. P.P. Pashinin, T.M. 23 Ushakova, E.E. Starchak, and E. M. Antipov, 2012. 24 18. M. Gall, P. Hendra, O. Peacock, M. Cudby and H. Willis, Spectrochimica Acta Part A: 25 Molecular Spectroscopy, 1972, 28, 1485-1496. 26 19. G. Strobl and W. Hagedorn, Journal of Polymer Science: Polymer Physics Edition, 1978, 16, 27 1181-1193. 28 29 20. M. Glotin, R. Domszy and L. Mandelkern, Journal of Polymer Science: Polymer Physics Edition, 30 1983, 21, 285-294. 31 21. M. Glotin and L. Mandelkern, Colloid Polym Sci, 1982, 260, 182-192. 32 22. K. Chernyshov, D. Gen, Y. Shemouratov, K. Prokhorov, G. Nikolaeva, E. Sagitova, P. Pashinin, 33

A. Kovalchuk, A. Klyamkina and P. Nedorezova, Macromolecular Symposia, 2010. Accepted 34 23. Y. V. Zavgorodnev, K. A. Prokhorov, L. A. Novokshonova, T. M. Ushakova, E. E. Starchak, V. G. 35 36 Krasheninnikov, G. Y. Nikolaeva, E. A. Sagitova, P. P. Pashinin, M. A. Guseva, P. Donfack, B. 37 von der Kammer and A. Materny, Laser Phys., 2012, 22, 730-737. 38 24. F. Rull, A. Prieto, J. Casado, F. Sobron and H. Edwards, Journal of Raman Spectroscopy, 1993, 39 24, 545-550. 40 25. B. Stuart, Vibrational spectroscopy, 1996, 10, 79-87. 41 26. A. G. Simanke, G. B. Galland, L. Freitas, J. A. H. da Jornada, R. Quijada and R. S. Mauler, 42 Polymer, 1999, 40, 5489-5495. 43 44 27. N. J. Everall, Applied Spectroscopy, 2000, 54, 773-782. 45 28. G. Leone, M. Mauri, S. Losio, F. Bertini, G. Ricci and L. Porri, Polymer Chemistry, 2014, 5, 46 3412-3423. Methods 47 29. M. Hong, L. Cui, S. Liu and Y. Li, Macromolecules, 2012, 45, 5397-5402. 48 49 50 51 52 53 54 55 56 57 58 59 Analytical 60