FTIR and molecular mechanics studies of H-bonds in aliphatic polyurethane and polyamide-66 model molecules

Available online at www.sciencedirect.com

Spectrochimica Acta Part A 69 (2008) 407–412

FTIR and molecular mechanics studies of H-bonds in aliphatic polyurethane and polyamide-66 model molecules Guoqing Wang a , Chunxia Zhang a , Xiaohe Guo b , Zhiyong Ren b,∗ a

School of Material and Chemical Engineering, Zhengzhou University of Light Industry, Zhengzhou 450002, China b Henan Key Laboratory of Fine Chemicals, Zhengzhou 450002, China Received 24 January 2007; received in revised form 3 April 2007; accepted 17 April 2007

Abstract Model aliphatic polyurethane (APU) hard segment based on 1,6-hexamethylene diisocyanate (HDI) and 1,4-butanediol (BDO) were prepared. FTIR and molecular mechanics (MM) simulation were used to conduct the systematic studies on APU and polyamide-66 (PA-66) whose sole difference lies in the alkoxyl oxygen. It was found that the introduction of the alkoxyl not only increases the conformations in APU, makes it a possible H-bond acceptor, but also weakens the H-bond between NH and O C in APU. There are two conformers stably existed in APU with lowest energy, leading to eight H-bond complexes based on NH as donor and (1) O C as acceptor, and another two complexes based on (2) alkoxyl O and (3) urethane N as acceptors, whereas there is only one stable conformer in PA-66, leading to one H-bond complex. One predominant H-bond complex has been found in APU with probability of about 95%. The simulated results are consistent with the νNH and νC O band shifting in FTIR. © 2007 Elsevier B.V. All rights reserved. Keywords: Polyurethane; Polyamide; H-bond; Molecular mechanics simulation

1. Introduction Segmented polyurethane (PU) and polyamide (PA) are both excellent polymer materials that have found many useful applications. These two types of polymers, which are both characterized by the H-bond between NH and O C, have been widely studied [1–27]. Included among them are various PU models [1–5] and segmented copolymers [6–13], as well as various PAs such as PA-5,10 [14], PA-11 [15], and PA-66 [16-17]. Most of the H-bonding interactions in these polymers were investigated by experimental methods while in recent years, molecular modeling or molecular modeling combined with FTIR has become a new way [18–21] in studying the H-bonds. Molecular mechanics (MM) simulation has been also approved to be an effective way especially in studying the effect of conformers on the H-bonds [22,23]. In addition, the study on H-bonding in PU has been focused on the relation with crystallization [9,24,25], mechanical properties [26]. The H-bonding interaction between PU and PA has been also investigated [27]. However, there is no report on the effect of alkoxyl on the H-bonds in PU.

∗

Corresponding author. Tel.: +86 371 65511795. E-mail address: [email protected] (Z. Ren).

1386-1425/$ – see front matter © 2007 Elsevier B.V. All rights reserved. doi:10.1016/j.saa.2007.04.014

If considering only the hard segment in PU, the structure unit of PU and PA is similar except the only remarkable difference: the alkoxyl oxygen in urethane group (Fig. 1). It is especially so if PA-66 was compared with an aliphatic PU (APU) based on HDI and BDO (Fig. 2). The H-bond in both polymers is closely related to their crystallization [9,24,25]. In addition, it is well known that the odd and even methylene numbers in PA have great influence on their melting point, which is just due to different H-bonding numbers with different number of methylene [27]. The PU based on HDI and BDO is similar to PA-66 except the additional alkoxyl oxygen in urethane group. Though there are numerous references dealing with H-bonds in PU, yet there are few reports concerning the alkoxyl oxygen in PU, especially the effect of the conformations on the H-bonding. Studying the effect of alkoxyl oxygen and the stable conformation on the Hbond in PU is of significance to understand the crystallization and properties in PA and PU. Molecular mechanics can be taken as an effectively simpler method compared with QM to obtain reasonable molecular structure and energy. Quality of MM calculation depends on molecular force field. COMPASS force field appeared recently as a high quality force field [28], and is expected to give its full play superiority in simulating different H-bond patterns with various configurations.

408

G. Wang et al. / Spectrochimica Acta Part A 69 (2008) 407–412

2.2. FTIR

Fig. 1. The characteristic group in APU and PA-66.

Therefore, in the present work, a model APU hard segment based on HDI and BDO was prepared. FTIR and molecular mechanics simulation were used to study the effect of alkoxyl oxygen on the H-bond in PU by making comparison with PA66 whose sole difference lies in the alkoxyl oxygen atom. The possible conformation and their H-bond complexes as well as the most probable H-bond complexes were also compared between APU and PA-66. 2. Experimental 2.1. Sample preparation APU was synthesized in a four-neck round bottom flask fitted with an overhead stirrer, an addition funnel, a thermometer and under a constant dry nitrogen blanket. The molar ratio of HDI and BDO was 1:1. The HDI dissolved in a solvent was introduced into the reaction vessel, and an equal molar of BDO dissolved in a solvent was introduced into the titration funnel. The chain extender solution was dropwise added to the diisocyanate solution with constant stirring at 80 ◦ C, and kept the temperature not higher than 90 ◦ C in the presence of dibutyltin dilaurate (DBTDL) as a catalyst. After the addition of BDO solution was over, the reaction was allowed to continue for one more hour between 80 and 90 ◦ C. The solvents were then evaporated, first at room temperature and then in a vacuum oven at a vacuum oven at 60 ◦ C for a week. PA-66, EPR27, produced and kindly supplied by China Shenma Engineering Plastics Co. Ltd., was used as received. The polymer exhibits a glass transition temperature of approximately 45 ◦ C and a crystalline melting point of 196 ◦ C, as determined by differential scanning calorimetry.

The infrared spectra as KBr pellets of APU powder was obtained using a SHIMADZU FTIR-8700 Spectrophotometer while the formic acid solution of PA-66 was directly coated onto a piece of KBr plate and measured after formic acid was completely removed. The frequency range covered was from 4000 to 400 cm−1 by averaging 32 scans at a resolution of 4 cm−1 . 2.3. Molecular modeling All molecular models were built on Silicon Graphics O2 workstation using the program Cerius 2 version 3.8 developed by Molecular Simulations Incorporated (MSI). COMPASS [28] force field was used to optimize each model molecule. The high convergence option was adopted for the energy minimization of each model; the root mean square of force on each atom was ˚ The model molecules controlled smaller than 0.001 kcal/mol/A. representing APU and PA-66 are N-methyl methyl carbamate and N-methyl acetamide in Fig. 3. For the individual molecule in a conformation, which can be a donor or an acceptor, the energy was obtained after being optimized; whereas for the H-bond complex, the two corresponding molecules were firstly located in a certain position where the H-bond donor atom, H and the acceptor atom are in line with a ˚ distance between the donor and the acceptor atom less than 3 A before optimization. The energy of both the individual donor and acceptor molecules, and the energy of the complex were then obtained. According to classical mechanics, the association energy between two single molecules includes all non-bonded molecular interactions such as the Lennard–Jones interactions, the Coulombic interactions and H-bond interaction. However, for the molecules with functional groups such as NH and C O in PU and PA, the energy resulting from the first two is usually one magnitude smaller than that from H-bond, hence it becomes a routine way to take the association energy, i.e. the difference between the sum energy of both individual molecules and the energy of the complex, E, as the H-bonding energy [18,20,21]. Hence in the present study, we also take such E as H-bonding energy Eh .

Fig. 2. APU and PA-66 structural units.

Fig. 3. Model molecules of APU and PA-66.

G. Wang et al. / Spectrochimica Acta Part A 69 (2008) 407–412

409

Fig. 4. Comparative FTIR spectra of APU and PA-66 in νNH and νC O regions.

3. Results and discussion

Fig. 5. Comparative FTIR spectra of APU and PA-66 in amide II and amide III.

3.1. Comparative FTIR spectra concerning H-bond of APU and PA-66

3.2. Molecular simulation

Fig. 4 is the comparative FTIR spectra of APU and PA66 at room temperature. Obvious wavenumber shifting can be found in APU and PU-66 samples, especially in νNH and νC O regions. The νNH band of APU is at 3321 cm−1 while that of PA-66 is at 3303 cm−1 ; the νC O band of APU is at 1786 cm−1 while that of PA-66 is at 1635 cm−1 , indicating both NH and C O of PA-66 are in the stronger H-bond state. It seems normal when both the donor and acceptor are either in the stronger or weaker H-bond state, i.e. both νNH and νC O are in the lower wavenumbers in PA-66 compared with that of APU. It is worth noting that both νNH and νC O may not be in the same shift trend in MDI and HMDI based PU [1]. When νNH of HMDI PU model segment is in the higher wavenumber, its νC O is in the lower wavenumber compared with MDI based PU model segment. Therefore, above results concerning the consistency of the wavenumber shifting in νNH and νC O can possibly only be seen in the relative simple H-bond interactions. In MDI and HMDI PU model hard segments, however, there may exist more factors affecting the wavenumber shifting, such as phenyl ring and cycloalkyl ring that may have different effects on H-bond patterns. In addition, the amide III bands (δ NH) in APU and PA66 are consistent with their νNH band. The amide III band at 1263 cm−1 in APU is in the lower wavenumber than that at 1277 cm−1 in PA-66, indicating from another respect that the NH in PA-66 is in the stronger H-bond. Yet we also note that the amide II bands in these two samples are exactly the same, both are at 1541 cm−1 (Fig. 5), suggesting amide III is more related to the H-bond based on NH and C O.

3.2.1. Model molecule selection and possible conformers in APU and PA-66 According to the structural unit of APU and PA-66, N-methyl methyl carbamate and N-methyl acetamide shown in Fig. 3 were selected to be their model molecules, respectively. The feasibility of the MM simulation on the similar structures have been tested elsewhere [22,23] and proved to be an effective way of modeling the optimization structure and energy difference. For APU, since there are two ␴-bonds in APU model molecule and each ␴-bond will lead to two states, therefore, there theoretically exist four different conformers like A, B, C and D by changing the position of the two methyls, which were shown in Fig. 6. The four letters A, B, C, D mean the four conformers. In energy optimization of the four conformers, we found that conformer D does not exist as it is so unstable that in the optimization turning to conformer B. Hence there possibly exist only three conformers, A, B, and C. However, since conformer C has much higher energy than conformers A and B (the energy difference is over 8 kcal/mol), the probability of conformer C can be taken hardly to exist. Therefore, conformers A and B can be actually taken as the models for further simulations on H-bonding interactions. For PA-66, however, since there is only one ␴-bond in PA model molecule, which leads to two states, there theoretically exist only two conformers shown in Fig. 7, two fewer than APU conformers. But only one may be taken as the model for further H-bonding interaction due to the much higher energy in another conformer. Therefore, two APU models (A and B in Fig. 6) and one PA-66 model (A in Fig. 7) will be used to model the H-bond interactions in APU and PA-66.

Fig. 6. Possible conformers in APU.

410

G. Wang et al. / Spectrochimica Acta Part A 69 (2008) 407–412

Fig. 7. Possible conformers in PA-66.

3.2.2. H-bond complexes based on different conformers in APU and PA-66 The two conformers (A and B in Fig. 6) of APU model molecules would result in four pairs of H-bonded complexes, including A + A, A + B, B + B, B + A. Since the H-bond donor in conformer A and B is different, H-bond complex (A + B) is different from that of (B + A) in forming H-bond. In addition, since there possibly exist the ‘pair conformation’ [23], the total num-

ber of the H-bond complexes will be doubled, leading to other additional four conformers like A + A , A + B , B + B , B + A . This is because the H-bonding related groups, NH and C O, are not in the central symmetric axis of the PU model molecule. Therefore, the formed H-bond pair (H-bond complex) resulted from the two conformers (e.g. A + A or A + B) thus possess different H-bond configurations when one of the two conformers in the pair rotates 180◦ around H-bond. Such two pairs of H-bond complexes are recognized as ‘pair conformation’. Only by considering the symmetry or conformations of the PU and PA, can we not lose the possible H-bond complexes. Then based on all these possible H-bond complexes, we can calculate each of the forming probability and get the most probable H-bond complex. Further, the probable H-bond can be correlated with the molecular structure. It is unsuitable to simply think the PU and PA should have some conformation without doing in this way, as some favorable conformers would be lost and the final H-bond complex may not be the reasonable one. Fig. 8 shows the whole eight possible H-bond complexes when pair conformation is considered. As is seen that, A + A or B + B is just the pair conformation of A + A or B + B. Table 1 lists their calculated H-bond properties. We note that in these H-bond com-

Fig. 8. Possible H-bond complexes in APU. Table 1 List of the H-bond angle, energy and distance in APU model molecule H-bond type

˚ H-bond O· · ·N (A)

H-bond angle O(N)· · ·H-N

ETotal (kcal/mol)

E(total) (kcal/mol)

σ (10−3 )

Prob. (%)

(A + A)I (A + A)I  (B + B)I (B + B)I  (A + B)I (A + B)I  (B + A)I (B + A)I  (A + A)II (A + A)III

2.897, (3.656/3.893) 2.906, (3.267/3.267) 2.911/2.911a 2.925, (3.280/3.280) 2.899, (3.522) 2.916/2.925 2.920, (3.275) 2.916, (3.207) 2.922/2.922a 3.402b , (3.456/3.811)

154.3, (121.0/103.6) 174.5, (111.7/101.3) 168.5/168.5 170.0, (104.8/106.4) 160.0, (133.4) 174.8/160.3 171.5, (106.7/104.6) 169.3, (104.4/107.3) 166.9/166.9 136.9, (112.9/118.9)

−91.520 −90.631 −94.170 −90.541 −90.351 −91.540 −90.839 −91.457 −89.164 −89.345

−7.792 −6.903 −5.389 −7.149 −6.791 −7.980 −7.279 −7.897 −2.886 5.953

12.07 2.74 1000 2.36 1.72 12.48 3.88 10.87 0.24 0.32

1.15 0.26 95.54 0.23 0.17 1.19 0.37 1.04 0.02 0.03

Note: data in bracket are weak H-bonds based on CH as donor. a Dimer. b Two different H-bonds coexisting.

G. Wang et al. / Spectrochimica Acta Part A 69 (2008) 407–412

411

Fig. 9. Possible H-bond complex in PA-66 (A + A). Fig. 11. Most probable H-bonds complexes in APU and PA-66.

plexes, there are two H-bond dimers. One is the H-bond dimer based on NH and O C, another is the H-bond dimer based on NH and alkoxyl. In PA-66 model molecule, however, there are expected to be only two H-bond complexes based on conformers (A + A) and (A + A ). Since the H-bond complexes based on (A + A ) turns to the exactly same H-bond complex as (A + A), there is actually only one H-bond complex to be considered in PA-66, as Fig. 9 shows, while there are three possible H-bond types in APU (Fig. 10) although the last two can only exist in the special configuration and the probability is very few. The more possible H-bond patterns in APU are expected to result from the alkoxyl in urethane. The only one type of H-bond complex in PA-66 is favorable for forming high regularity of the structure with high crystallinity. 3.2.3. The most probable H-bond complexes in APU and PA-66 Probability of above H-bond complexes can be calculated, as we did before [22,23], according to Boltzmann equation shown as follows: σ(i) = e−Ec (i)/RT  e−Ec (i)/RT Z=

(1) (2)

i

P(i) =

σ(i) Z

(3)

In order to calculate the probability of each H-bonded complex, we firstly calculate the statistical weight factor σ (i) for each configuration by using Eq. (1), then calculate the partition function Z by using Eq. (2), and finally obtain the probability P(i) for each H-bonded complex by Eq. (3). Here Ec (i) is

the energy difference between the energy in each H-bond complex and the H-bond complex with the lowest energy. Here R is 1.987 × 10−3 kcal/mol, T is taken 300 K (at about room temperature), hence RT is simply taken as 0.6 in order to make easier calculation. It can be seen from Table 1 that in addition to the eight conformers based on NH as donor and C O as acceptor, other two H-bond conformers based on NH as donor, and alkoxyl as well as nitrogen in urethane as acceptors are also included in this Table. Therefore, there are three possible types of H-bonds in APU. The first type comprises eight different conformers, known as I after the letters brackets standing for the conformer type, such as (A + A)I, (B + B)I, etc.; the second type is the H-bond based on NH as donor and alkoxyl as acceptor, known as II after the letters brackets standing for the conformer type such as (A + A)II; the third type is the H-bond based on NH as donor and urethane N atom as acceptor, known as III after the letters brackets standing for the conformer type such as (A + A)III. The optimization results between the conformers A and B show that only conformer A may form the type II and type III H-bond. It is found from Table 1 that the H-bond conformer (B + B)I based on NH· · ·O C H-bond shown in left side of Fig. 11 (dimer) is predominant, as its probability amounts to almost 95% among all the H-bonded conformers. The probabilities of other seven possible complexes are among 0.17–1.19. The probability of type II and III H-bonds are only 0.02 and 0.03%, though simulation shows they may exist. In another respect, the predominant H-bond complex in PA-66 model is expected to be the only model complex shown in the right side of Fig. 11. It can be seen by making the comparison of the most probability between APU and PA-66 model molecules that the H-bond distance (O· · ·N) of PA-66 is shorter than that of APU and the

Fig. 10. Possible H-bond types in APU based on NH as donor.

412

G. Wang et al. / Spectrochimica Acta Part A 69 (2008) 407–412

Table 2 Average H-bond energy and length of three types of H-bonds in both PA-66 and APU model molecules

lower than that of APU, which is consistent with FTIR result, suggesting the H-bond in PA-66 is stronger than that of APU.

H-bond type

PA-66 NH· · ·O C

APU NH· · ·O C

APU NH· · ·OCO

Acknowledgements

˚ Average H-bond length (A) H-bond energy (kcal/mol)

2.896 −7.655

2.911 −5.389

2.922 −2.886

This study has been supported by the National Nature Science Foundation (20474014) and Henan Nature Science Foundation (0611020700). We also appreciate professor Yang for his help on simulation conducted at his laboratory in Institute of Chemistry, Chinese Academy of Sciences.

H-bond energy of PA-66 is lower (−7.655 kcal/mol) than that of APU (−5.389 kcal/mol), which is reasonable and consistent with FTIR results. 3.2.4. Comparative H-bond energy and length between APU and PA-66 model molecules Table 2 lists the H-bond properties in both APU and PA66. Here we list the properties of only three H-bond complexes including the two types of H-bonds in APU, as they are both dimers and the energy calculation is relatively more accurate. The third type of H-bond based on NH as donor and the urethane N as acceptor comprises other week H-bond based on CH and C O, therefore the energy cannot be the pure. The week H-bond is not the main subject in the present paper, therefore it will be discussed in more details in our future work. In a word, the results from FTIR and MM are consistent each other, while the study of comparative H-bonds in APU and PA66 model molecules clearly shows the effect of alkoxyl oxygen on the H-bonding in them. The introduction of alkoxyl into the model molecule increases not only the conformations, but also the H-bond type. In addition, it affects the H-bonding properties such as length and energy. The H-bond NH· · ·O C with alkoxyl has longer length and higher energy than that free of alkoxyl. 4. Conclusion The introduction of alkoxyl into the model molecule to become urethane increases not only the conformations, but also the H-bond type. There are theoretically two conformers in PA66 model molecule but only one can exist stably; while there are four conformers in APU model molecule and two of them can stably exist, leading to eight H-bond complexes based on NH and O C, and leading to other two H-bond complexes based on NH and O–C or urethane N. The H-bond energy in PA-66 is

References [1] S. Zhang, Z. Ren, S. He, Y. Zhu, C. Zhu, Spectrochim. Acta 66 (2007) 188. [2] Z. Ren, H. Wu, J. Ma, D. Ma, Chin. J. Polym. Sci. 22 (2004) 225. [3] M.M. Coleman, H.L. Keun, D.J. Skrovanek, P.C. Painter, Macromolecules 19 (1986) 2149. [4] C.M. Brunette, S.L. Hsu, W.J. MacKnight, Macromolecules 15 (1982) 71. [5] C.S.P. Sung, N.S. Schneider, Macromolecules 8 (1975) 68. [6] J.P. Sheth, D.B. Klinedinst, G.L. Wilkes, I. Yilgor, E. Yilgor, Polymer 46 (2005) 7317. [7] N. Luo, D.-N. Wang, S.-K. Ying, Macromolecules 30 (1997) 4405. [8] A.M. Fernandez, A.E. Lozano, L. Gonzales, A. Rodriguez, Macromolecules 30 (1997) 3584. [9] N. Luo, D.-N. Wang, S.-K. Ying, Polymer 37 (1996) 3577. [10] M.M. Coleman, P.C. Painter, Prog. Polym. Sci. 20 (1995) 1. [11] T. Yamamoto, M. Shibayama, S. Nomura, Polym. J. 21 (1989) 895. [12] C.S.P. Sung, N.S. Schneider, Macromolecules 10 (1977) 452. [13] R.W. Seymour, G.M. Estes, S.L. Cooper, Macromolecules 3 (1970) 579. [14] P. Villasenor, L. Franco, J.A. Subirana, J. Puiggali, J. Polym. Sci. Part B: Polym. Phys. 37 (1999) 2383. [15] D.J. Skrovanek, P.C. Painter, M.M. Coleman, Macromolecules 19 (1986) 699. [16] D. Garcia, H.W. Starkweather Jr., J. Polym. Sci., Polym. Phys. Ed. 23 (1985) 537. [17] D.J. Skrovanek, S.E. Howe, P.C. Painter, M.M. Coleman, Macromolecules 18 (1985) 1676. [18] E. Yilg¨or, I. Yilgor, E. Yurtsever, Polymer 43 (2002) 6551. [19] V.L. Furer, J. Mol. Struct. 520 (2000) 117. [20] H. Sun, Macromolecules 26 (1993) 5924. [21] J. Bandekar, S. Klima, Spectrochim. Acta 48 (1992) 1363. [22] Z. Ren, X. Zeng, X. Yang, D. Ma, S.L. Hsu, Polymer 46 (2005) 12337. [23] Z. Ren, D. Ma, X. Yang, Polymer 44 (2003) 6419. [24] E. Yilg¨or, E. Yurtsever, I. Yilg¨or, Polymer 43 (2002) 6561. [25] R.L. McKiernan, A.M. Heintz, S.L. Hsu, E.D.T. Atkins, J. Penelle, S.P. Gido, Macromolecules 35 (2002) 6970. [26] Y.I. Tien, K.H. Wei, Polymer 42 (2001) 3213. [27] A. Genovese, R.A. Shanks, Comput. Theor. Polym. Sci. 11 (2001) 57. [28] H. Sun, J. Phys. Chem. B 102 (1998) 7338.