Separation of maleic anhydride grafted polypropylene using multidimensional high-temperature liquid chromatography

G Model

ARTICLE IN PRESS

CHROMA-357355; No. of Pages 10

Journal of Chromatography A, xxx (2016) xxx–xxx

Contents lists available at ScienceDirect

Journal of Chromatography A journal homepage: www.elsevier.com/locate/chroma

Separation of Maleic Anhydride grafted Polypropylene using Multidimensional High-Temperature Liquid Chromatography K.N. Prabhu a , T. Macko a , R. Brüll a,∗ , K. Remerie b , J. Tacx b , P. Garg b , A. Ginzburg b a Fraunhofer Institute for Structural Durability and System Reliability, Division Plastics, Group Material Analytics, Schlossgartenstrasse 6, 64289 Darmstadt, Germany b SABIC Technology & Innovation, STC Geleen, P.O. Box 319, 6160 AH Geleen, The Netherlands

a r t i c l e

i n f o

Article history: Received 17 December 2015 Received in revised form 5 February 2016 Accepted 29 February 2016 Available online xxx Keywords: Functionalized polypropylene liquid chromatography HT-HPLC HT-SEC HT 2D-LC IR-detection

a b s t r a c t Functionalization addresses a property gap of polyolefins and opens new perspectives due to improved surface properties in applications like composites (e.g., glass fiber reinforced polypropylene) and anticorrosive coatings for metals. Various techniques have been developed to characterize functionalized polyolefins, yet no analytical approach addressing their chemical heterogeneity exists. Using High Temperature Size Exclusion Chromatography (HT-SEC) coupled to infrared spectroscopy we could show for two model samples of polypropylene grafted maleic anhydride (PP-g-MA), differing in their nominal MA content, that the grafting density increases with decreasing molar mass. Crystallization Analysis Fractionation (CRYSTAF) does not enable to separate these samples according to their composition to the extent required. Yet, when using High Temperature High Performance Liquid Chromatography (HT-HPLC), with either silica gel or Mica as stationary phase and a gradient mobile phase, a deformulation into a grafted and a non-grafted fraction could be achieved. This was confirmed by analyzing the eluted fractions by infrared spectroscopy. Hyphenating the separation according to composition with a separation according to molar mass (HT-HPLC x HT-SEC) enabled for the first time to reveal the bivariate distribution of PP-g-MA with regard to the molar mass and composition. Using on-line infrared detection quantitative information on the compositional and molar mass parameters of the individual fractions could be obtained. © 2016 Published by Elsevier B.V.

1. Introduction Polyolefins are the most extensively used synthetic polymers due to their widely adaptable application properties and production from inexpensive feedstocks. Currently they constitute, by volume, in excess of 60% of all synthetic polymers produced [1]. Nevertheless, their low surface energy and poor compatibility with other (polar) polymers impose a limit on certain applications. In the same sense, their adhesion to materials like wood, metals or reinforcing fibers requires special attention [2]. Introducing polar functionalities is one of the solutions to these problems, which can be achieved by grafting suitable polar monomers onto the polymer chain. The chemical modification of polypropylene using reactive extrusion has been an area of intense interest and the grafting of maleic anhydride (MA) on polypropy-

∗ Corresponding author. Tel.: +49 6151 705 8639; fax: +49 6151 7058601. E-mail address: [email protected] (R. Brüll).

lene (PP) is of high commercial relevance [3] for which the state of the art has been well reviewed [1,4–7]. The application properties of the products of reactive extrusion of PP and MA are, for a given overall composition, determined by their molar mass distribution (MMD) and chemical composition distribution (CCD). The MMD and the corresponding average values can be determined by high temperature size exclusion chromatography (HT-SEC). The average chemical composition of such reaction products can be analyzed by spectroscopic techniques. FTIR spectroscopy [8–11] has been widely used for this purpose due to its sensitivity and ease of measurement. The strengths of nuclear magnetic resonance spectroscopy are structure elucidation and the ability to deliver compositional information without prior calibration [12–15]. Titration techniques [8,11,16–19] are the classical approach to determine the average content of MA-groups bound on polyolefins. However, all these techniques deliver average values for the degree of functionalization, and no information about the molecular heterogeneities i.e., the way the comonomer is distributed along and across the molar mass axis, is obtained. For the case

http://dx.doi.org/10.1016/j.chroma.2016.02.081 0021-9673/© 2016 Published by Elsevier B.V.

Please cite this article in press as: K.N. Prabhu, et al., Separation of Maleic Anhydride grafted Polypropylene using Multidimensional High-Temperature Liquid Chromatography, J. Chromatogr. A (2016), http://dx.doi.org/10.1016/j.chroma.2016.02.081

G Model CHROMA-357355; No. of Pages 10

ARTICLE IN PRESS K.N. Prabhu et al. / J. Chromatogr. A xxx (2016) xxx–xxx

2

of polypropylene grafted maleic anhydride (PP-g-MA) extraction methods have been investigated to separate the grafted product from unreacted PP [11,20]. However, such techniques are not very selective, and serious drawbacks from an industrial standpoint are low sample throughput and poor reproducibility. This creates the need for an analytical technique which can separate these materials according to their degree of functionalization. For deformulation of olefin copolymers, approaches based on crystallization have been widely used: Temperature Rising Elution Fractionation (TREF) [21], Crystallization Analysis Fractionation (CRYSTAF) or Crystallization Elution Fractionation (CEF) [22–24] separate semi-crystalline polyolefins according to their comonomer content. These analytical techniques use the fact that the crystallization temperature of polyolefins from dilute solution is related to their comonomer content. When applying crystallization based techniques on PP-g-MA the dependency of the crystallization on the molar mass [25,26] of the polymer is the main problem. Furthermore, the graft content is typically very low (< 5 mol %), and the crystallization temperature, which is the variant for the separation, depends also on the microstructure (with regard to stereo- and regio-chemistry) of the grafted units on the backbone [27]. Liquid chromatography has recently shown potential to separate polymers bearing polar groups. Thus, statistical copolymers of ethylene with various polar monomers [28–30] could be chromatographically separated with regard to their composition. These separations were based on the full adsorption of macromolecules from specific solvents on bare silica gel as stationary phase and a subsequent desorption of the macromolecules by applying a gradient to a thermodynamically good solvent. An advantage of such a chromatographic separation for PP-g-MA would be the possibility of hyphenation with a separation according to molar mass, which then gives access to the bivariate distribution (CCD x MMD). Interestingly, none of these modern analytical techniques has yet been applied to post reactor functionalized PP. In this paper we want to probe the suitability of crystallization based techniques and liquid chromatography to determine the bivariate distribution of PP-g-MA. 2. Experimental

Fig. 1. Temperature profile of preparative CRYSTAF.

Fig. 2. Schematic setup of HT-SEC-LC Transform-FTIR.

Isotactic PP (i-PP) with a weight average molar mass, Mw , as 95 kg/mol (i-PP95 ) was obtained from American Polymer Standards Corp. (Mentor, USA). Polystyrene (PS) standards (Agilent, Germany) with an Mw of 580 − 6780000 g/mol and Ð of 1.03 - 1.12 were used to calibrate the HT-SEC. The polymer solutions were prepared by dissolving the samples in an adsorption promoting solvent (decalin) at concentrations of about 1–1.5 mg/mL. The samples were dissolved at 140 ◦ C and the time for dissolution was in the range of 2–3 h.

2.1. Column packings Silica gel (PerfectSil® 300 Å from MZ Analysentechnik, Mainz, Germany) with a particle diameter of 5 ␮m was packed into a column of 250 × 4.6 mm L. x I.D. Particles of Mica (muscovite) from Creations Couleurs, The Innovation Company, Dreux, France, with a diameter of 5–15 ␮m, were dry packed manually into a column of 150 × 4.6 mm L. x I.D. The composition of muscovite may be expressed by the molecular formula KAl2 (AlSi3 O10 )(F,OH)2 and this mineral exhibits hexagonal morphology. A PLgel Olexis column, 300 × 7.5 mm L. x I.D., (Polymer Laboratories, Church Stretton, England) containing particles of 10 ␮m diameter was used for the HT-SEC measurements. 2.2. Solvents and Polymer Samples Decalin and cyclohexanone were used as received. 1,2,4trichlorobenzene (TCB) and 1,2-dichlorobenzene (ODCB) were distilled prior to use. The solvents were obtained from Merck, Darmstadt, Germany. Two PP-g-MA samples were obtained from SABIC (Geleen, The Netherlands). Their molecular characterization data are detailed in Table 1.

2.3. Crystallization Analysis Fractionation A CRYSTAF [22] apparatus, model 200, manufactured by Polymer Char S.A. (Valencia, Spain) was used. About 20 mg of the sample were dissolved in 30 mL of ODCB at 160 ◦ C. After dissolution the temperature of the sample solution was decreased at a rate of 0.1 ◦ C/min from 100 ◦ C to 30 ◦ C. The polymer concentration in solution was monitored by an IR detector operating at 160 ◦ C and using 3.5 ␮m as the measuring wavelength. A Prep mc2 instrument (Polymer Char, Valencia, Spain) was used to collect fractions from CRYSTAF. 500 mg of polymer dissolved in 300 mL ODCB were separated into 3 fractions with different elution temperatures (Fig. 1). 2.4. Coupling HT-SEC with FTIR spectroscopy via LC Transform interface Fig. 2 schematically shows the procedure followed: The polymer sample of interest was separated using a PL GPC 220 (Polymer Laboratories, Church Stretton, England). Three PLgel Olexis columns were used with TCB as the mobile phase at a flow rate of 0.5 mL/min. The entire system was thermostated at 150 ◦ C. After HT-SEC fractionation the analyte was deposited on a rotating germanium disc

Please cite this article in press as: K.N. Prabhu, et al., Separation of Maleic Anhydride grafted Polypropylene using Multidimensional High-Temperature Liquid Chromatography, J. Chromatogr. A (2016), http://dx.doi.org/10.1016/j.chroma.2016.02.081

G Model CHROMA-357355; No. of Pages 10

ARTICLE IN PRESS K.N. Prabhu et al. / J. Chromatogr. A xxx (2016) xxx–xxx

3

Table 1 Characteristics of the samples. Sample 95

i-PP PP-g-MA PP-g-MA 1 2

Average amount of polar monomer [wt. %]1

Weight average molar mass, Mw [g/mol]2

Dispersity, Ð

n.a. 0.4 0.7

95000 145000 112000

3.4 5.1 3.4

Determined by 1 H-NMR at SABIC (Geleen, The Netherlands). Calculated from HT-SEC calibration with PS standards. n.a.-not applicable.

in the LC Transform solvent evaporation interface (Series 300, Lab Connections, Marlborough, MA, USA). The rotating speed of the germanium disc was 10◦ /min. The deposited trace of the polymer was analyzed off-line by FTIR spectroscopy (NicoletTM iSTM 10 FTIR spectrometer, Thermo Scientific, Dreieich, Germany) at a scan rate of 1◦ /spectrum. 16 scans were accumulated per spectrum and a baseline was automatically subtracted for each measurement. Gram-Schmidt plots [31–33] and the profiles of the carbonyl index (CI) were calculated from the obtained data sets. The GramSchmidt plot reflects the overall integrated intensity of the IR absorption (2800–3000 cm−1 ) along the chromatographic run. The CI is defined as the ratio of the carbonyl absorption (range 1700–1780 cm−1 ) to the −CH2 - and −CH3 - vibration (2800–3000 cm−1 ) of the polymer.

2.5. High Temperature Liquid Chromatography Chromatographic measurements were conducted with a PL GPC 120 high-temperature liquid chromatograph (Polymer Laboratories, Church Stretton, England), which included a robotic sample handling system (PL-XTR 220) for injection of sample solutions. A sample loop of 100 ␮L was used. A constant flow rate of 0.8 mL/min and a constant temperature (140 ◦ C) were maintained for all experiments. PerfectSil® or a column filled with Mica particles were used as a sorbent. Binary mobile phases of varying composition were generated by a high-pressure gradient pump (Agilent, Waldbronn, Germany) operating at room temperature. The mobile phase was heated to 140 ◦ C in a transfer capillary before entering the injector. After the injection of a polymer solution the adsorption promoting liquid (decalin) was pumped for 3 min, then the pump started to deliver the desorption promoting liquid (cyclohexanone) in a linear gradient from 0 to 100% cyclohexanone in 10 min. Finally, the column was purged with decalin for 23 min (> 10 times the column dead volume) with the aim to re-establish the initial conditions for the next injection. The polymer samples were fully soluble also in the desorption promoting solvent (cyclohexanone). An evaporative light scattering detector (ELSD), model PLELS 1000 (Polymer Laboratories, Church Stretton, England), was used. The nebulizer temperature and the evaporation temperature were 160 ◦ C and 260 ◦ C, respectively, with a nitrogen flow rate of 1.5 L/min. Data collection and processing were done using WinGPC software from Polymer Standards Services, Mainz, Germany.

area of the peaks was used to determine the degree of crystallinity and the CI. 2.7. High-temperature two-dimensional liquid chromatography The measurements were realized with a high-temperature twodimensional liquid chromatograph (PolymerChar, Valencia, Spain), comprising an autosampler, two ovens, valves and two pumps equipped with vacuum degassers (Agilent, Waldbronn, Germany). One oven was used for thermostating the HT-SEC column (PLgel Olexis column, 300 × 7.5 mm L × I.D.). The second one, where the injector and switching valves were housed, was used to thermostat the HT-HPLC column (PerfectSil® ). The sample solution was filtered through an in-built stainless steel filter in the HT 2D-LC instrument. This filter was automatically flushed back after each filtration. Moreover, stainless steel frits with a pore size of 2 ␮m were part of the PerfectSil® column. A scheme of the setup is shown in Fig. 3. The HT-HPLC and HT-SEC columns were hyphenated by an electronically controlled eight-port valve system EC8W (VICI Valco instruments, Houston, Texas, USA) equipped with two 100 ␮L loops. First dimension separations were carried out on a PerfectSil® column (HT-HPLC), and a PLgel Olexis column was used in the second dimension (HT-SEC). From the moment of injection into the HT-HPLC column (100 ␮L injection loop) the 8-port valve was switched every 5 minutes in order to inject 100 ␮L of effluent from the HT-HPLC into the HTSEC column. The flow rate in the HT-HPLC column was 0.02 mL/min, while that of TCB in the HT-SEC column was 1.5 mL/min. The separations were done by applying a linear gradient decalin → cyclohexanone. Starting with 100% of decalin for 200 min, the concentration of cyclohexanone in the mobile phase was linearly raised to 100% within 700 min and then held constant for 100 min. Finally, the column was purged with decalin (30 min at 1 mL/min), until the initial chromatographic conditions were reestablished. The complete 2D-LC analysis required ∼1000 minutes. The effluent from the HT-SEC column was monitored by an IR4 detector (Polymer Char, Valencia, Spain). The concentration of polymer in the effluent was monitored by a broadband IR filter centered at 2900 cm−1 , which monitors the absorbance due to C-H bonds in macromolecules. The ovens, the autosampler, as well as all transfer lines were thermostated at 140 ◦ C. The parameters in the 2D-LC instrument were handled by software provided by PolymerChar (Valencia, Spain). Software WinGPC 7.0 (Polymer Standards Service, Mainz, Germany) was used for data acquisition and evaluation.

2.6. Preparative HT-HPLC → FTIR 3. Results and Discussion The polymer eluting in the mobile phase before and after starting the gradient was collected 20 times separately in two 20 mL vials (i.e., in total 4.0 mg of a sample were fractionated). The solvent in the bottles was evaporated at 100 ◦ C in vacuo, and the obtained polymer was analyzed by FTIR spectroscopy. The IR spectra of the polymer samples were recorded using a Nicolet 8700 instrument (Thermo Fisher Scientific, Dreieich, Germany) with an MCT-detector at a spectral resolution of 4 cm−1 in ATR mode. 16 scans were accumulated per spectrum. The half

Two samples of PP-g-MA, which differ in their average content of MA, were chosen and analyzed by HT-SEC → FTIR using the LCTransform approach [28,34,35]. The results are presented in Fig. 4. Fig. 4 reveals that the extent of the grafting is indirectly proportional to the molar mass of PP-g-MA i.e., a gradient of functionalization exists along the MMD. It is supposed that during the reactive extrusion used for preparation of these samples the iPP molecules undergo chain scission and ␤-elimination, wherein

Please cite this article in press as: K.N. Prabhu, et al., Separation of Maleic Anhydride grafted Polypropylene using Multidimensional High-Temperature Liquid Chromatography, J. Chromatogr. A (2016), http://dx.doi.org/10.1016/j.chroma.2016.02.081

G Model CHROMA-357355; No. of Pages 10 4

ARTICLE IN PRESS K.N. Prabhu et al. / J. Chromatogr. A xxx (2016) xxx–xxx

Fig. 3. Setup for HT 2D-LC [29,30].

Fig. 4. Gram-Schmidt plot and CI against molar mass for a) PP-g-MA0.4 and b) PP-g-MA0.7 . Notice: HT-SEC axis was calibrated with PS standards. The superscript in PP-g-MA denotes the average MA content in the samples (Table 1).

chain scission leads to a reduction in molar mass. Nevertheless, the generation of free radicals during chain scission increases the extent of the grafting reactions. Thus low and medium molar mass components display more grafting compared high molar mass fractions. From these results the average degree of functionalization along the MM axis can be extracted, while fractions of same molar mass but differing in their degree of functionalization will co-elute, which then necessitates a deformulation with regard to CC. To study the compositional distribution perpendicular to the MM axis (CCD) crystallization techniques from solution have been widely used for semi-crystalline olefin copolymers [22,36]. The separation in these techniques is based on the crystallization of macromolecules from a hot dilute solution in a temperature gradient. On the basis of Flory’s theory [37] the crystallization temperature of a random copolymer is explicitly related to its comonomer content. To probe the suitability of such an approach the PP-g-MA samples were analyzed by CRYSTAF, together with an i-PP sample as reference (Fig. 5). From CRYSTAF a profile of concentration versus temperature is obtained and its first derivative (dW/dT) delivers information about the CCD (Fig. 5). The area under the cumulative curve, which denotes the fraction crystallizing in a respective temperature range, is listed in Table 2.

Fig. 5. Overlay of the first derivatives of the polymer concentration in solution, dW/dT, of i-PP and PP-g-MA samples.

Fig. 5 shows that all samples crystallize with a sharp peak between 85 and 70 ◦ C. Tc of about 79 ◦ C, 77 ◦ C and 75 ◦ C for i-PP95 , PP-g-MA0.4 , PP-g-MA0.7 respectively are evident. The homopolymer i-PP exhibits a small amount of amorphous fraction (Table 2).

Please cite this article in press as: K.N. Prabhu, et al., Separation of Maleic Anhydride grafted Polypropylene using Multidimensional High-Temperature Liquid Chromatography, J. Chromatogr. A (2016), http://dx.doi.org/10.1016/j.chroma.2016.02.081

G Model

ARTICLE IN PRESS

CHROMA-357355; No. of Pages 10

K.N. Prabhu et al. / J. Chromatogr. A xxx (2016) xxx–xxx

5

Table 2 Area under the curve and peak crystallization temperature, Tc , for the fractions. Sample

i-PP95 PP-g-MA0.4 PP-g-MA0.7 *

*

Tc [◦ C]

79 77 75

Area between two temperatures [%] Fraction 1 (85−70 ◦ C)

Fraction 2 (70−30 ◦ C)

Fraction 3 (< 30 ◦ C)

94 54 56

3 20 15

3 26 29

Tc − Crystallization temperature at peak maxima.

Fig. 6. Overlay of the IR spectra of PP-g-MA0.4 fractions in the wavenumber range a) 940–1020 cm−1 and b) 1660–1820 cm−1 and the bulk sample (PP-g-MA0.4 ) in the region c) 940–1020 cm−1 and d) 1660–1820 cm−1 .

This can be explained by the fact that the i-PP standard contains PP of lower isotacticity as confirmed by NMR. Both PP-g-MA samples exhibit a tailing from 70 to 45 ◦ C, which can be seen to be more pronounced for PP-g-MA0.4 . A significant amount of amorphous fraction, which does not crystallize at 30 ◦ C, is observed for both samples (Fig. 5 and Table 2). It is assumed that this loss in crystallinity is the result of the grafting. Taking into account that the crystallization temperature is related to the degree of function-

alization [28] it may then be derived that both PP-g-MA samples display considerable heterogeneity with regard to their functionalization (Fraction 2 in Table 2). To further substantiate this CRYSTAF fractions were collected for PP-g-MA0.4 in the temperature ranges as specified in Table 2 and Fig. 1 and analyzed by FTIR-spectroscopy. The results are presented in Fig. 6 and the calculated degree of crystallinity and the CI are compiled in Table 3.

Table 3 Band area (A) ratio corresponding to crystallinity and CI for all three fractions of PP-g-MA0.4 . Fractions

1 2 3

A998 /A973 (Degree of crystallinity)

0.74 0.73 0.70

CI (A of MA bands/A973 ) A1710 /A973

A1740 /A973

A1780 /A973

0.06 0.15 n.f

0.01 0.17 n.f

0.07 0.09 0.16

*n.f.- not feasible as the peaks merged with other characterisitic peaks.

Please cite this article in press as: K.N. Prabhu, et al., Separation of Maleic Anhydride grafted Polypropylene using Multidimensional High-Temperature Liquid Chromatography, J. Chromatogr. A (2016), http://dx.doi.org/10.1016/j.chroma.2016.02.081

G Model CHROMA-357355; No. of Pages 10 6

ARTICLE IN PRESS K.N. Prabhu et al. / J. Chromatogr. A xxx (2016) xxx–xxx

Fig. 7. Overlay of chromatograms of i-PP95 and PP-g-MA samples using a) PerfectSil® and b) Mica at 140 ◦ C. Notice: Composition of the mobile phase at the ELSD is indicated in the figures.

Fig. 8. Overlay of the IR spectra corresponding to PP-g-MA0.4 and to the fractions eluting before and in the gradient in the wavenumber range a) 950-1020 cm−1 and b) 1550–1850 cm−1 .

Fig. 9. Overlay of HT-SEC traces recorded by HT 2D-LC → IR a) before the gradient (i-PP) and b) in the gradient (PP-g-MA). (HT-HPLC: 0.02 mL/min and HT-SEC: 1.5 mL/min), 100 ␮L transfer loop.

The three fractions as well as the bulk sample show absorptions characteristic for i-PP at 973 cm−1 (representing the amorphous/crystalline phase) and at 998 cm−1 (representing the crystalline phase), and their area ratio (A998 /A973 ) denotes the degree of crystallinity [38] (Fig. 6a and c). All three fractions vary in their degree of crystallinity (Table 3) and, as can be expected from the sequence of crystallization, fraction 3 is less crystalline compared to the other ones.

The bulk sample and all three fractions show absorptions in the carbonyl region (Fig. 6b and d). Three bands (1710, 1740 and 1780 cm−1 ) were observed in the carbonyl region of the bulk sample (PP-g-MA0.4 ) (Fig. 6b): The bands at 1710 and 1740 cm−1 signify the carbonyl vibration of carboxylic acid and ester respectively [10] and the band at 1780 cm−1 can be attributed to the symmetric stretching of the carbonyl groups in the MA-ring [8,18,19]. The finding of the acid and the ester group in the bulk sample may

Please cite this article in press as: K.N. Prabhu, et al., Separation of Maleic Anhydride grafted Polypropylene using Multidimensional High-Temperature Liquid Chromatography, J. Chromatogr. A (2016), http://dx.doi.org/10.1016/j.chroma.2016.02.081

G Model CHROMA-357355; No. of Pages 10

ARTICLE IN PRESS K.N. Prabhu et al. / J. Chromatogr. A xxx (2016) xxx–xxx

be the result of the presence of either a different type of grafting [18] or the open chain acid anhydride [39]. It can be noticed that the absorbance at 1780 cm−1 is in all three fractions more intense compared to the bulk sample, while that at 1740 cm−1 is lower (Fig. 6b and d). A plausible explanation would be that the acid group reverts back to the anhydride during the experimentation at high temperature. Fraction 1 and 2 (Fig. 6b) show similar characteristic bands as the bulk (Fig. 6c). Thus it can be concluded that the first fraction has a lower MA content compared to the second one (Table 3 and Fig. 6b). CRYSTAF enables to separate PP-g-MA according to the degree of functionalization, but the resolution is not sufficient to achieve a full deformulation into a grafted and non-grafted fraction. This provides the motivation to develop HT-HPLC as an alternative method for the CCD determination. Albrecht et al. [28,34] investigated polar sorbents and mobile phases for separating statistical copolymers of ethylene with various polar monomers. Graft copolymers contain polar groups on a non-polar backbone and it can, therefore, be assumed that the polar groups will selectively interact with a polar sorbent, while non-polar units will not contribute to the retention of the macromolecules. Accordingly, Mica and silica gel (PerfectSil® ) were selected as sorbents. PerfectSil® has been applied previously for the analysis of random copolymers of ethylene with polar comonomers [28,40], while to the best of our knowledge no applications of Mica for the chromatography of polymers have been reported. Mica is a layered silicate with hexagonal morphology and this arrangement leads to its thermal stability of up to 500 ◦ C. Decalin and cyclohexanone were found to be an adequate adsorption and desorption promoting solvent, respectively [28]. The elugrams of PP-g-MA samples using the chromatographic systems PerfectSil® /decalin → cyclohexanone and Mica/decalin → cyclohexanone are shown in Fig. 7. Both functionalized samples eluted in two peaks in both sorbents (Fig. 7): The first one, eluting before the gradient, represents a portion of the sample which was not adsorbed in the column, while the second one is adsorbed on the sorbent and later desorbed after addition of cyclohexanone to the mobile phase. As a rule the retention of a polar analyte increases with its polarity, and consequently it may be assumed that the latter eluting fraction can be assigned to PP grafted with MA. To verify this hypothesis the effluent of PP-g-MA0.4 corresponding to both peaks was collected for the system PerfectSil® /decalin → cyclohexanone and analyzed by FTIR spectroscopy (Fig. 8). It can be recognized that the polymer eluting before as well as in the gradient shows vibrations at 998 and 973 cm−1 , which are indicative of i-PP (Fig. 8a). The polymer eluting in the gradient shows carbonyl vibrations similar to those of the bulk, which are not present in the fraction eluting before the gradient (Fig. 8b). Thus, from Fig. 7 and Fig. 8b it can be confirmed that HT-HPLC separates the reaction product into a grafted (PP-g-MA) and non-grafted fraction (i-PP). For a meaningful interpretation of chromatographic data, a quantitative detector response is essential: While the ELSD remains the sole option for one dimensional HT-HPLC its response is exponential with regard to concentration and depends on the composition of the eluting polymer [41]. Moreover, to derive structure↔property relationships from analytical results knowledge about the bivariate distribution with regard to composition and molar mass is imperative. Two dimensional liquid chromatography (HT-HPLC x HT-SEC), which hyphenates the compositional separation with one according to molar mass, has proven to be appropriate for this purpose [29,42,43]. An additional advantage of adding the HT-SEC dimension is the fact that in SEC the elution occurs quasi isocratic, which then enables to employ IR spec-

7

Fig. 10. Response of the IR-detector with respect to concentration of the injected sample solution with i-PP95 .

troscopy for detection [44,45]. Since the IR detector is tuned to the stretching vibration of the C-H bonds a baseline separation of the solvent plug (cyclohexanone + decalin/TCB) from the transferred fraction of polymer is mandatory. This was achieved by using a HT-SEC column with 11000 theoretical plates and by providing sufficient time for the individual HT-SEC analyses [41]. Representative HT-SEC traces using the optimized conditions are shown in Fig. 9. From Fig. 9 it was observed that both HT-HPLC fractions (iPP and PP-g-MA) are baseline separated from the intense solvent peaks. The IR response was calibrated (Fig. 10) by injecting solutions of i-PP95 with varying concentrations into the HT 2D-LC → IR. Both PP-g-MA samples were analyzed by HT 2D-LC → IR using the above described optimized experimental conditions and the results are presented as two dimensional contour plots with respective projections in Fig. 11. The complete 2D-LC analysis required ∼1000 minutes, however, the time needed for realization of 2D-LC may be significantly shortened after modifying several parameters, like it was outlined in our previous works [29,41,42]. The contour plot of both samples reveals two spots: The one which elutes before the gradient can be assigned to the homopolymer, i-PP, and the other, which elutes in the gradient, to PP-g-MA. The broadness of the first peak on the HT-HPLC axis (i.e., polymer eluting in decalin is separated via SEC mechanism in the HPLC column) reflects the MMD of i-PP, while the broadness of the second peak on the same axis reflects the CCD of the PP-g-MA. Differences in the HT-HPLC elution volume (Fig. 7 and Fig. 11) are due to different instrumental setups. To determine the MM and MMD of both constituents the contour plots were projected into the SEC plane (Fig. 12). Fig. 12 is obtained when the results from HT-2D-LC → IR are projected into the SEC dimension and Fig. 4 arises from HT-SEC → FTIR using the LC-Transform approach. Although the origin of the data is different it can be noted that the MMDs are comparable with regard to their range (103 − 106 g/mol) and shape. Using the calibration (Fig. 10) the portion of i-PP present in the analyzed samples (PP-g-MA0.4 ) may be calculated from the peak areas of i-PP eluting before the gradient (Fig. 11a). By subtracting the amount of i-PP (eluting before the gradient) from the total amount of the injected polymer, the amount of PP-g-MA (eluting in the gradient) was obtained. The results are summarized in Table 4. It is evident from Fig. 12 that both samples exhibit a broad MMD (1 × 103 to 1 × 106 g/mol) and contain significant amounts of low molar mass material in the PP-g-MA part (1 × 103 to 1 × 104 g/mol). Table 4 reveals that the Mw for the homopolymer (i-PP) is higher than that of the graft copolymer (PP-g-MA). The HT-SEC axis was

Please cite this article in press as: K.N. Prabhu, et al., Separation of Maleic Anhydride grafted Polypropylene using Multidimensional High-Temperature Liquid Chromatography, J. Chromatogr. A (2016), http://dx.doi.org/10.1016/j.chroma.2016.02.081

G Model CHROMA-357355; No. of Pages 10

ARTICLE IN PRESS K.N. Prabhu et al. / J. Chromatogr. A xxx (2016) xxx–xxx

8

Fig. 11. 2D contour plot of a) PP-g-MA0.4 and b) PP-g-MA0.7 . Notice: The HT-SEC axis was calibrated with PS standards.

calibrated with PS standards. PS standards are regularly used for calibration of SEC separations, as they are widely available with low dispersities over a wide molar mass region. However, the hydrodynamic volume of i-PP can be different from that of PP-g-MA for chains having the same absolute molar mass, due to intramolecular interactions. In the case that the incorporation of polar groups significantly decreases the hydrodynamic volume, the apparent molar

mass would be lower. The fact that the sample with higher MA content (0.7 wt. %) shows lower masses for the grafted part (Fig. 12 and Table 4), compared to the sample with lower MA content (0.4 wt. %), is in agreement with this hypothesis. Furthermore both samples contain ∼20 wt. % of i-PP homopolymer and their portion of grafted copolymer is similar (∼80 wt. %) (Table 4).

Table 4 Composition of the analyzed samples as calculated from the peak areas in thechromatograms. Sample

i-PP [wt. %]

Mw of i-PP [g/mol]

PP-g-MA [wt. %] (calculated from wt. % of i-PP)

Mw of PP-g-MA [g/mol]

PP-g-MA0.4 PP-g-MA0.7

20 18

2.3 × 105 8.2 × 104

80 82

1.1 × 105 4.2 × 104

Please cite this article in press as: K.N. Prabhu, et al., Separation of Maleic Anhydride grafted Polypropylene using Multidimensional High-Temperature Liquid Chromatography, J. Chromatogr. A (2016), http://dx.doi.org/10.1016/j.chroma.2016.02.081

G Model CHROMA-357355; No. of Pages 10

ARTICLE IN PRESS K.N. Prabhu et al. / J. Chromatogr. A xxx (2016) xxx–xxx

9

Fig. 12. Overlay of the HT-SEC projections from HT 2D-LC → IR corresponding to a) PP-g-MA0.4 and b) PP-g-MA0.7 .

4. Conclusion Despite the fact that various analytical techniques have been applied in the past to characterize functionalized polyolefins, the challenge of determining the bivariate distribution (MMD and CCD) of reaction products generated from polypropylene and maleic anhydride remained unsolved. By using HT-SEC → FTIR via the LC-Transform interface we could show that the grafting density increases with decreasing molar mass. Although a compositional separation could be achieved via CRYSTAF, the selectivity of such a crystallization based approach is not sufficient for PP-g-MA. Yet, using HT-HPLC with silica gel (PerfectSil® ) and Mica as stationary phase and a solvent gradient decalin → cyclohexanoneG–1 min at 140 ◦ C, two samples of PP-g-MA were baseline separated into a functionalized and a non-functionalized portion. The separation achieved was confirmed by analyzing the HT-HPLC fractions (obtained with PerfectSil® as a sorbent) with FTIR spectroscopy. This separation according to the chemical composition was hyphenated with a separation according to the molar mass, which enabled for the first time to determine the bivariate distribution of PP-g-MA samples. The obtained contour plots from HT 2D-LC → IR exhibited a two spot regime, reflecting the grafted and non-grafted component. From the contour plots it could be shown that the two samples, which differ in their nominal content of maleic anhydride are comparable with regard to their relative content of the grafted material. Yet, a higher degree of grafting is accompanied by a lower molar mass of the grafted portion. The developed analytical methodology may be highly useful for developing more efficient processes of functionalization, and the analytical information can be applied to derive structure↔property relationships for functionalized polyolefins. Conflict of interest The authors declare no competing financial interest. Acknowledgements This research forms part of the research collaboration between the group Material Analytics, Division Plastics, Fraunhofer Institute, LBF, Darmstadt, and SABIC Technology and Innovation (T & I), SABIC Technology Centre (STC), Geleen, The Netherlands. The authors acknowledge Mr. Subin Damodaran and Mr. Jan-Hendrik Arndt for the FTIR and the LC Transform measurements and interpretation respectively. The authors would also like to acknowledge Dr.

Frank Malz (Fraunhofer Institute, LBF, Darmstadt), Dr. Mark Boerakker (SABIC, Geleen) and Dr. Haika Hildebrandt (SABIC, Geleen) for all their inputs. We would also like to thank Ms. Laetitia Lourtil, Dr. Sébastien Thumerel and Dr. Alain Saintrond from Creations Couleurs, The Innovation Company, France for providing us the Mica material and related information. References [1] T.C. Chung, Functionalization of polyolefins, first ed., Academic Press California, USA, 2002. [2] T.C. Chung, Synthesis of functional polyolefin copolymers with graft and block structures, Prog. Polym. Sci. 27 (2002) 39–85, http://dx.doi.org/10.1016/ S0079-6700(01)00038-7. [3] M.D. Purgett, O. Vogl, Functional polymers. XLVIII. Polymerization of ␻-alkenoate derivatives, J. Polym. Sci. Part A: Polym. Chem. 26 (1988) 677–700, http://dx.doi.org/10.1002/pola.1988.080260302. [4] M.W.R. Brown, A.F. Johnson, P.D. Coates, Reactive processing of polymers, Rapra Technology Limited, first ed., Smithers Rapra Publishing, UK, 1994. [5] V. Mittal, Advances in polyolefin nanocomposites, first ed., CRC Press, USA, 2010. [6] Z. Mengjiao, C. Fulin, C. Lan, Z. Yanhao, Research progress on preparation and application of maleic anhydride grafted polymer, China Elastomerics 3 (2013) 17. [7] Z.M. Rzayev, Graft copolymers of maleic anhydride and its isostructural analogues: High performance engineering materials, Int Rev. Chem. Eng. 3 (2011) 153–215, arXiv: 1105.1260. [8] M. Sclavons, M. Laurent, J. Devaux, V. Carlier, Maleic anhydride-grafted polypropylene: FTIR study of a model polymer grafted by ene-reaction, Polymer 46 (2005) 8062–8067, http://dx.doi.org/10.1016/j.polymer.2005.06. 115. [9] S.H.P. Bettini, J.A.M. Agnelli, Grafting of maleic anhydride onto polypropylene by reactive extrusion, J. Appl. Polym. Sci. 85 (2002) 2706–2717, http://dx.doi. org/10.1002/app.10705. [10] M. Sclavons, P. Franquinet, V. Carlier, G. Verfaillie, I. Fallais, R. Legras, M. Laurent, F.C. Thyrion, Quantification of the maleic anhydride grafted onto polypropylene by chemical and viscosimetric titrations, and FTIR spectroscopy, Polymer 41 (2000) 1989–1999, http://dx.doi.org/10.1016/ S0032-3861(99)00377-8. [11] S.H.P. Bettini, J.A.M. Agnelli, Evaluation of methods used for analyzing maleic anhydride grafted onto polypropylene by reactive processing, Polym. Test. 19 (2000) 3–15, http://dx.doi.org/10.1016/S0142-9418(98)00066-X. [12] W. Heinen, M. van Duin, C.H. Rosenmöller, C.B. Wenzel, H.J.M. De Groot, J. Lugtenburg, 13 C NMR Study of the grafting of 13 C labeled maleic anhydride onto PE, PP and EPM, J. Macromol. Symp. 129 (1998) 119–125, http://dx.doi. org/10.1002/masy.19981290111. [13] L. Yang, F. Zhang, T. Endo, T. Hirotsu, Microstructure of maleic anhydride grafted polyethylene by high-resolution solution-state NMR and FTIR spectroscopy, Macromolecules 36 (2003) 4709–4718, http://dx.doi.org/10. 1021/ma020527r. [14] Q. Cheng, Z. Lü, H.J. Byrne, Synthesis of a maleic anhydride grafted polypropylene–butadiene copolymer and its application in polypropylene/styrene–butadiene–styrene triblock copolymer/organophilic montmorillonite composites as a compatibilizer, J. Appl. Polym.Sci. 114 (2009) 1820–1827, http://dx.doi.org/10.1002/app.30678. [15] M. Zhang, R.H. Colby, S.T. Milner, T.C.M. Chung, T. Huang, W. deGroot, Synthesis and characterization of maleic anhydride grafted polypropylene

Please cite this article in press as: K.N. Prabhu, et al., Separation of Maleic Anhydride grafted Polypropylene using Multidimensional High-Temperature Liquid Chromatography, J. Chromatogr. A (2016), http://dx.doi.org/10.1016/j.chroma.2016.02.081

G Model CHROMA-357355; No. of Pages 10 10

[16]

[17]

[18]

[19]

[20]

[21]

[22]

[23]

[24]

[25]

[26]

[27]

[28]

[29]

[30]

[31]

ARTICLE IN PRESS K.N. Prabhu et al. / J. Chromatogr. A xxx (2016) xxx–xxx

with a well-defined molecular structure, Macromolecules 46 (2013) 4313–4323, http://dx.doi.org/10.1021/ma4006632. Y. Minoura, M. Ueda, S. Mizunuma, M. Oba, The reaction of polypropylene with maleic anhydride, J. Appl. Polym. Sci. 13 (1969) 1625–1640, http://dx. doi.org/10.1002/app.1969.070130805. N.G. Gaylord, M.K. Mishra, Nondegradative reaction of maleic anhydride and molten polypropylene in the presence of peroxides, J. Polym. Sci.: Polym. Lett. Ed. 21 (1983) 23–30, http://dx.doi.org/10.1002/pol.1983.130210104. B. De Roover, M. Sclavons, V. Carlier, J. Devaux, R. Legras, A. Momtaz, Molecular characterization of maleic anhydride-functionalized polypropylene, J. Polym. Sci. Part A: Polym. Chem. 33 (1995) 829–842, http:// dx.doi.org/10.1002/pola.1995.080330509. M. Sclavons, V. Carlier, B. De Roover, P. Franquinet, J. Devaux, R. Legras, The anhydride content of some commercial PP-g-MA: FTIR and titration, J. Appl. Polym. Sci. 62 (1996) 1205–1210, http://dx.doi.org/10.1002/(SICI) 10974628(19961121) 62:8<1205:AID-APP10>3.0.CO;2-6. M. Fujiyama, T. Wakino, H. Wachi, K. Tani, Studies on grafting of maleic-anhydride onto polypropylene in molten state by use of screw extruder, Kobunshi Ronbunshu 49 (1992) 97–104. L. Wild, T.R. Ryle, D.C. Knobeloch, I.R. Peat, Determination of branching distributions in polyethylene and ethylene copolymers, J. Polym. Sci.: Polym. Phys. Ed. 20 (1982) 441–455, http://dx.doi.org/10.1002/pol.1982.180200307. B. Monrabal, Encyclopedia of Analytical Chemistry, in: R.A. Mayers (Ed.), TREF and CRYSTAF technologies for polymer characterization, John Wiley & Sons, Inc., New York, 2000, pp. 1–20. B. Monrabal, J. Sancho-Tello, N. Mayo, L. Romero, Crystallization elution fractionation. A new separation process for polyolefin resins, Macromol. Symp. 257 (2007) 71–79, http://dx.doi.org/10.1002/masy.200751106. B. Monrabal, N. Mayo, R. Cong, Crystallization elution fractionation and thermal gradient interaction chromatography. Techniques comparison, Macromol. Symp. 312 (2012) 115–129, http://dx.doi.org/10.1002/masy. 201100017. S. Anantawaraskul, J.B.P. Soares, P.M. Wood-Adams, Cocrystallization of blends of ethylene/1-olefin copolymers: An investigation with crystallization analysis fractionation (CRYSTAF), Macromol. Chem. Phys. 205 (2004) 771–777, http://dx.doi.org/10.1002/macp.200300116. K. Suriya, S. Anantawaraskul, J.B.P. Soares, Cocrystallization of ethylene/1-octene copolymer blends during crystallization analysis fractionation and crystallization elution fractionation, J. Polym. Sci. Part B: Polym. Phys. 49 (2011) 678–684, http://dx.doi.org/10.1002/polb.22231. T. Macko, U. Schulze, R. Brüll, A. Albrecht, H. Pasch, T. Fónagy, L. Häussler, B. Iván, Monitoring the Chemical Heterogeneity and the Crystallization Behavior of PP-g-PS Graft Copolymers Using SEC-FTIR and CRYSTAF, Macromol. Chem. Phys. 209 (2008) 404–409, http://dx.doi.org/10.1002/macp.200700398. A. Albrecht, R. Brüll, T. Macko, F. Malz, H. Pasch, Comparison of high-temperature HPLC, CRYSTAF and TREF for the analysis of the chemical composition distribution of ethylene-vinyl acetate copolymers, Macromol. Chem. Phys. 210 (2009) 1319–1330, http://dx.doi.org/10.1002/macp. 200900135. A. Ginzburg, T. Macko, V. Dolle, R. Brüll, High-temperature two-dimensional liquid chromatography of ethylene-vinylacetate copolymers, J. Chromatogr. A 1217 (2010) 6867–6874, http://dx.doi.org/10.1016/j.chroma.2010.08.045. A. Ginzburg, T. Macko, F. Malz, M. Schroers, I. Troetsch-Schaller, J. Strittmatter, R. Brüll, Characterization of functionalized polyolefins by high-temperature two-dimensional liquid chromatography, J. Chromatogr. A 1285 (2013) 40–47, http://dx.doi.org/10.1016/j.chroma.2013.01.067. S. de Goede, R. Brüll, H. Pasch, N. Marshall, Monitoring thermo-oxidative degradation of polypropylene by CRYSTAF and SEC-FTIR, Macromol. Symp. 193 (2003) 35–44, http://dx.doi.org/10.1002/masy.200390062.

[32] H. Pasch, M.I. Malik, Advanced Separation Techniques for Polyolefins, first ed., Springer International Publishing, Switzerland, 2014, http://dx.doi.org/10. 1016/j.progpolymsci.2013.10.005. [33] Å. Björck, Numerics of Gram-Schmidt orthogonalization, Linear Algebra Appl. 197 (1994) 297–316, http://dx.doi.org/10.1016/0024-3795(94) 90493-6. [34] A. Albrecht, R. Brüll, T. Macko, H. Pasch, Separation of ethylene-vinyl acetate copolymers by high-temperature gradient liquid chromatography, Macromolecules 40 (2007) 5545–5551, http://dx.doi.org/10.1021/ ma070732c. [35] S. Cheruthazhekatt, T.F. Pijpers, G.W. Harding, V.B. Mathot, H. Pasch, Multidimensional analysis of the complex composition of impact polypropylene copolymers: Combination of TREF, SEC-FTIR-HPer DSC, and high temperature 2D-LC, Macromolecules 45 (2012) 2025–2034, http://dx. doi.org/10.1021/ma2026989. [36] B. Monrabal, J. Blanco, J. Nieto, J.B.P. Soares, Characterization of homogeneous ethylene/1-octene copolymers made with a single-site catalyst. CRYSTAF analysis and calibration, J. Polym. Sci. Part A: Polym. Chem. 37 (1999) 89–93, http://dx.doi.org/10.1002/(SICI) 1099-0518(19990101) 37:1<89:AIDPOLA10>3.0.CO;2-#. [37] P.J. Flory, Thermodynamics of crystallization in high polymers. iv. A theory of crystalline states and fusion in polymers, copolymers, and their mixtures with diluents, J.Chem. Phys. 17 (1949) 223–240, http://dx.doi.org/10.1063/1. 1747230. [38] D.R. Burfield, P.S.T. Loi, The use of infrared spectroscopy for determination of polypropylene stereoregularity, J. Appl. Polym. Sci. 36 (1988) 279–293, http:// dx.doi.org/10.1002/app.1988.070360203. [39] I. Noda, A. Dowrey, J. Haynes, C. Marcott, Group frequency assignments for major infrared bands observed in common synthetic polymers, in: James E. Mark (Ed.), Physical properties of polymers handbook, Springer, USA, 2007, pp. 395–406. [40] A. Albrecht, R. Brüll, T. Macko, P. Sinha, H. Pasch, Analysing the chemical composition distribution of ethylene-acrylate copolymers: comparison of HT-HPLC, CRYSTAF and TREF, Macromol. Chem. Phys. 209 (2008) 1909–1919, http://dx.doi.org/10.1002/macp.200800223. [41] K.N. Prabhu, R. Brüll, T. Macko, K. Remerie, J. Tacx, P. Garg, A. Ginzburg, Separation of bimodal high density polyethylene using multidimensional high temperature liquid chromatography, J. Chromatogr. A 1419 (2015) 67–80, http://dx.doi.org/10.1016/j.chroma.2015.09.078. [42] T. Macko, A. Ginzburg, K. Remerie, R. Bruell, Separation of High-Impact Polypropylene Using Interactive Liquid Chromatography, Macromol. Chem. Phys. 213 (2012) 937–944, http://dx.doi.org/10.1002/macp.201200017. [43] A. Roy, M.D. Miller, D.M. Meunier, A.W. Degroot, W.L. Winniford, F.A. Van Damme, R.J. Pell, J.W. Lyons, Development of comprehensive two-dimensional high temperature liquid chromatography × gel permeation chromatography for characterization of polyolefins, Macromolecules 43 (2010) 3710–3720, http://dx.doi.org/10.1021/ma100010e. [44] A. Ortín, J. Montesinos, E. López, P. del Hierro, B. Monrabal, J.R. Torres-Lapasió, M.C. García-Álvarez-Coque, Characterization of chemical composition along the molar mass distribution in polyolefin copolymers by GPC using a modern filter-based IR detector, Macromol. Symp. 330 (2013) 63–80, http://dx.doi. org/10.1002/masy.201300060. [45] B. Monrabal, P. del Hierro, Characterization of polypropylene-polyethylene blends by temperature rising elution and crystallization analysis fractionation, Anal. Bioanal. Chem. 399 (2011) 1557–1561, http://dx.doi.org/ 10.1007/s00216-010-4061-5.

Please cite this article in press as: K.N. Prabhu, et al., Separation of Maleic Anhydride grafted Polypropylene using Multidimensional High-Temperature Liquid Chromatography, J. Chromatogr. A (2016), http://dx.doi.org/10.1016/j.chroma.2016.02.081