- Email: [email protected]
nitrene chemistry
G Model
ARTICLE IN PRESS
APSUSC-34309; No. of Pages 9
Applied Surface Science xxx (2016) xxx–xxx
Contents lists available at ScienceDirect
Applied Surface Science journal homepage: www.elsevier.com/locate/apsusc
Full length article
Photografting of perfluoroalkanes onto polyethylene surfaces via azide/nitrene chemistry Konstantin Siegmann a , Jan Inauen a , Diego Villamaina b , Martin Winkler a,∗ a Institute of Materials and Process Engineering (IMPE), School of Engineering (SoE), Zurich University of Applied Sciences (ZHAW), Technikumstrasse 9, CH-8401 Winterthur, Switzerland b Visiting Scientist at IMPE, Permanent Address: Rapidplatz 3, CH-8953 Dietikon, Switzerland
a r t i c l e
i n f o
Article history: Received 15 July 2016 Received in revised form 31 October 2016 Accepted 1 November 2016 Available online xxx Keywords: Nitrene Azide Insertion Polyethylene Photografting Perfluorinated compounds
a b s t r a c t Water, oil and dirt repellent surfaces of plastic materials are desirable for many applications. Polyethylene is a common plastic and, because of its inertness, difficult to graft. We chose polyethylene as example because of its ubiquity and model character. As graft chains perfluoroalkanes were chosen, and photografting was selected as grafting method. Photolytically generated nitrenes can insert into carbon-hydrogen bonds and are therefore suited for binding to polyethylene. Hydrophobic photo reactive surface modifiers based on azide/nitrene chemistry are designed, synthesized in high yield and characterized. Four new molecules are described. One problem is to demonstrate that the photografted surface modifiers are bound covalently to the polyethylene. Abrasion tests show that all new molecules, when photografted to polyethylene, have a higher abrasion resistance than a polyethylene surface coated with a long-chain perfluoroalkane. An abrasion model using ice is developed. Also, other issues are addressed, such as differences in the UVspectra of the new compounds in solution and in the neat state. Exitonic coupling of the chromophores of the surface modifiers is observed for specific molecules. A linear correlation of water contact angle with fluorine surface content, as measured by photoelectron spectroscopy, in grafted polyethylene surfaces is established. © 2016 Published by Elsevier B.V.
1. Introduction Highly hydrophobic surfaces are desirable for many purposes, such as water, oil and dirt repellent outfit of polymeric materials. Polyethylene finds widespread uses as commodity plastic in a wealth of consumer goods. We therefore set ourselves to modify polyethylene surfaces in such a way that they become strongly hydrophobic. Further, this hydrophobicity should be resistant against abrasion and wear. From the point of view of a chemist, polyethylene is interesting as an example polymer, because it consists of carbon–carbon and carbon-hydrogen bonds only, which make it very inert and difficult to modify. If polyethylene surfaces can be rendered strongly and permanently hydrophobic, then other plastics could presumably be treated in the same way. Perfluorinated alky chains are known to possess a high water- and
∗ Corresponding author. E-mail addresses: [email protected] (K. Siegmann), [email protected] (J. Inauen), [email protected] (D. Villamaina), [email protected] (M. Winkler).
oil repellency which makes them apt to be used as hydrophobic surface modifiers. The task therefore is to graft polyethylene with perfluorinated alky chains. The modification of polymer surfaces has become an important theme for many applications [1]. From the methods developed, surface grafting has emerged as simple, useful and versatile approach. One of the major techniques is ultra violet (UV) light induced grafting, in all of its variations [2]. Yet unreactive surfaces, such as the surfaces of polyolefines, can be grafted using UV-light [3]. However, drastic conditions have to be used which were originally developed for photoaffinity labeling in bio-chemistry [4]. This method relies on photo chemically generated singlet nitrenes. A singlet nitrene is a nitrogen species having only six valence electrons (a sextet), that is highly reactive and therefore capable of inserting even into carbon-hydrogen bonds of the substrate [5,6]. This approach has been applied to surfaces, too. Two publications describe the surface modification of polysulfon with polyethylene glycol using photo chemically generated nitrenes for the prevention of cell adhesion [7,8]. The precursor molecule for the generation of a nitrene species usually is an azide [9,10]. Organic azides intermediately form
http://dx.doi.org/10.1016/j.apsusc.2016.11.007 0169-4332/© 2016 Published by Elsevier B.V.
Please cite this article in press as: K. Siegmann, et al., Photografting of perfluoroalkanes onto polyethylene surfaces via azide/nitrene chemistry, Appl. Surf. Sci. (2016), http://dx.doi.org/10.1016/j.apsusc.2016.11.007
G Model APSUSC-34309; No. of Pages 9
ARTICLE IN PRESS K. Siegmann et al. / Applied Surface Science xxx (2016) xxx–xxx
2
nitrenes by the uptake of energy, such as temperature or light, followed by the elimination of dinitrogen. However, because nitrenes are highly reactive, rearrangements of the parent molecule are frequent [11]. In order to convert azides into nitrenes by UV-radiation, the azide has to be bound to a chromophore (a light absorbing part), that is, an aromatic moiety. Since aromatic systems are stabilized by resonance, rearrangements of the intermediate nitrene are somewhat suppressed. However, one of the main reactions of phenyl nitrenes is their ring expansion to azacycloheptatetraenes which react further and do not yield the insertion product [12]. Hence, perfluorinated phenyl nitrenes were investigated and it was found that by their use rearrangements can be hindered in such a way that the desired carbon-hydrogen insertion reaction becomes dominant [12,13]. Adhesion promotors on the basis of photoreactive perfluorophenyl azide derivatives have been developed [14]. The desired hydrophobic and oleophobic effect is caused by the highly water repellent perfluorinated carbon chain. Such a chain has to be attached to the perfluorophenyl azide in order to obtain a photo reactive surface modifier. As fluorotelomer alcohols are readily available (they are extensively employed as water and dirt repellant substituents in the textile industry) it immediately suggests itself to use those alcohols as the hydrophobic part in the new molecules. Linking the fluorotelomer alcohol to the perfluorophenyl azide moiety is conveniently achieved by an ester functionality. The target molecule is therefore composed of four segments: The azide as the reactive group, the perfluorinated aromatic ring as chromophore and stabilizator, the ester functionality as linker and the fluorotelomer alcohol as the hydrophobic part. In the following sections synthesis, photochemistry and stability studies of the new molecules and surfaces are described.
2. Materials and methods 2.1. Materials All Chemicals were used without further purification. Fluorocarbon chemicals were purchased from Apollo Scientific Ltd, UK. Anhydrous dichloromethane was purchased from TCI Deutschland GmbH. Sodium azide (purum, p. a., ≥99.0%) and triethylamine (puriss, p. a., ≥99.5%) were from Sigma-Aldrich Chemie GmbH, Switzerland. Other chemicals were purchased from Sigma-Aldrich Chemie GmbH, Switzerland. Polyethylene was a high-density Borstar ME3440 type from Borealis, and the polyethylene plates had a thickness of 2 mm.
2.2. Instrumentation Melting points were determined on a Differential Scanning Calorimeter, DSC 204 F1 Phoenix from Netzsch Gerätebau GmbH. UV-Spectra were recorded on a Perkin Elmer 950 Spectrophotometer, equipped with a 150 mm integrating sphere for the neat spectra; fluorescence spectra on a LS 55 Fluorescence Spectrometer from Perkin Elmer. ATR-IR spectra were recorded using a Perkin Elmer Spectrum 100 FT-IR Spectrometer. Most NMR Spectra were run on a Bruker AscendTM 500 instrument. The MALDI-MS and the elemental analyses were from the Lab. for Organic Chemistry, ETH Zurich, Switzerland. Static water contact angles were measured using a Krüss DSA 100 instrument. XPS Spectra were obtained using a SPECSTM spectrometer from SPECS GmbH, Berlin, Germany. The UV radiometer was an UVpad spectral radiometer from Opsytec Dr. Gröbel GmbH, Germany. The UV belt drier was from Uviterno AG, Switzerland, equipped with a 150 mm light tube with a maximal power of 3 kW. Standard irradiation procedure was as follows: Speed of the belt: 0.026 m/s; lamp power 1.2 kW; dis-
tance lamp-sample 0.07 m. The spray-coated polyethylene samples were exposed 2 to 3 times to these conditions. 2.3. General method for the synthesis of pentafluorobenzoates (4a–d) The reaction was kept under an inert atmosphere of dry nitrogen at all times. To a stirred solution of 1 equiv. fluorotelomer alcohol 3a–d in dry dichloromethane (about 0.3 molar for alcohols 3a–c; about 0.05 molar for alcohol 3d) at 0 ◦ C was added 1.05 equiv. perfluorobenzoyl chloride (2) via a syringe. Then, 2.1 equiv. of triethylamine were added via a syringe. The mixture was stirred for 2 h at 0 ◦ C, allowed to warm to room temperature and stirred overnight. Work up: The solution was washed twice with equal amounts of 1 molar hydrochloric acid. Then, the organic phase was washed once with saturated sodium bicarbonate solution and once with water. The organic phase was dried with anhydrous sodium sulfate, filtered, and the dichloromethane was distilled off. The perfluorobenzoates were not purified but used as obtained for the next step. 2.4. 1H, 1H, 2H, 2H-Nonafluoro-1-hexyl Pentafluorobenzoate (4a) Liquid, yield 95% ATR-IR 1746, 1653, 1499, 1331, 1216, 1132, 1007 cm−1 ; 1 H NMR (500 MHz, CDCl3 ) ␦ 4.70 (t, J = 6.4 Hz, 2H) 2.63 (t x t, J = 6.4 Hz, 3 J(19 F, 1 H) = 18.1 Hz, 2H); 13 C NMR (126 MHz, CDCl3 ) ␦ 158.6, 146.7–136.6, 121–107, 58.3 t, 3 J(19 F, 13 C) = 4.6 Hz, 30.3 (t, 2 J(19 F, 13 C) = 22.0 Hz); 19 F NMR (471 MHz, CDCl3 ) ␦ −81.4 (m), −114.1 (m), −124.7 (m), −126.2 (m), −138.1 (m), −148 (m), −160.5 (m). 2.5. 1H, 1H, 2H, 2H-Tridecafluoro-1-octyl Pentafluorobenzoate (4b) Liquid, yield 96% ATR-IR 1746, 1653, 1500, 1331, 1190, 1144, 1009 cm−1 ; 1 H NMR (500 MHz, CDCl3 ) ␦ 4.70 (t, J = 6.5 Hz, 2H) 2.63 (t x t, J = 6.4 Hz, 3 J(19 F, 1 H) = 18.1 Hz, 2H); 13 C NMR (126 MHz, CDCl3 ) ␦ 158.6, 146.7–136.6, 121–107, 58.2 (t, 3 J(19 F, 13 C) = 4.5 Hz), 30.3 (t, 2 J(19 F, 13 C) = 21.8 Hz); 19 F NMR (471 MHz, CDCl3 ) ␦ −81.5 (m), −114.1 (m), −122.3 (m), −123.3 (m) −126.2 (m), −124.0 (m), −126.6 (m), −138.4 (m), −148.5 (m), −161.0 (m). 2.6. 1H, 1H, 2H, 2H-Heptadecafluoro-1-decyl Pentafluorobenzoate (4c) Solid, yield 97% ATR-IR 1738, 1652, 1501, 1332, 1197, 1145, 987 cm−1 ; 1 H NMR (300 MHz, CDCl3 ) ␦ 4.69 (t, J = 6.4 Hz, 2H) 2.61 (t x t, J = 6.4 Hz, 3 J(19 F, 1 H) = 18.0 Hz, 2H); 13 C NMR (75 MHz, CDCl3 ) ␦ 158.6, 147–136, 119–107, 58.3 (t, 3 J(19 F, 13 C) = 4.8 Hz), 30.4 (t, 2 J(19 F, 13 C) = 22.0 Hz); 19 F NMR (282 MHz, CDCl ) ␦ −81.1 (m), 3 −113.8 (m), −121.8 (m), −122.1 (m), −122.9 (m), −123.7 (m), −126.3 (m), −138.0 (m), −147.9 (m), −160.4 (m). 2.7. 1H, 1H, 2H, 2H Henicosafluoro-1-dodecyl Pentafluorobenzoate (4d) Solid, yield 97% ATR-IR 1738, 1653, 1529, 1497, 1199, 1147, 1006 cm−1 ; 1 H NMR (500 MHz, CDCl3 ) ␦ 4.68 (t, J = 6.4 Hz, 2H) 2.61 (t x t, J = 6.4 Hz, 3 J(19 F, 1 H) = 18.1 Hz, 2H); 13 C NMR (126 MHz, CDCl3 ) ␦ 158.8, 147–137, 119–107, 58.5 (t, 3 J(19 F, 13 C) = 4.6 Hz), 30.6 (t, 2 J(19 F, 13 C) = 21.9 Hz); 19 F NMR (471 MHz, CDCl3 ) ␦ −81.0 (m), −113.8 (m), −121.9 (m), −122.0 (m), −122.8 (m), −123.7 (m), −126.3 (m), −137.9 (m), −147.7 (m), −160.3 (m).
Please cite this article in press as: K. Siegmann, et al., Photografting of perfluoroalkanes onto polyethylene surfaces via azide/nitrene chemistry, Appl. Surf. Sci. (2016), http://dx.doi.org/10.1016/j.apsusc.2016.11.007
G Model
ARTICLE IN PRESS
APSUSC-34309; No. of Pages 9
K. Siegmann et al. / Applied Surface Science xxx (2016) xxx–xxx
F
Solid, yield 95%, recrystallized from ethanol, yield 74%, mp. 91 ◦ C, ATR-IR 2139, 1733, 1644, 1491, 1198, 1148, 997 cm−1 ; 1 H NMR (500 MHz, CDCl3 ) ␦ 4.67 (t, J = 6.4 Hz, 2H) 2.61 (t x t, J = 6.5 Hz, 3 J(19 F, 1 H) = 18.0 Hz, 2H); 13 C NMR (126 MHz, CDCl ) ␦ 159.0, 3 146.7–139.5, 124–107, 58.2 (t, 3 J(19 F, 13 C) = 4.5 Hz), 30.5 (t, 2 J(19 F, 13 C) = 21.7 Hz); 19 F NMR (471 MHz, CDCl ) ␦ −80.8 (m), −113.7 3 (m), −121.8 (m), −122.8 (m), −123.6 (m), −126.2 (m), −138.3 (m), −150.8 (m). Anal. Calcd for C19 H4 F25 N3 O2 : C, 29.21; H, 0.52; N, 5.38. Found: C, 29.12; H, 0.49; N, 5.52.
3. Results and discussion 3.1. Synthesis As described in the Introduction section, a photoreactive surface modifier consists of four segments. These can be seen in Fig. 1: the azide as reactive group, the perfluororophenyl moiety as chromophore, the ester functionality as linker and a fluorotelomer alcohol as the hydrophobic part. Four different fluorotelomer residues, RF , were investigated, and they are assigned to the letters a, b, c, and d throughout the paper. The fluorotelomer residues are non-branched, linear chains with an even number of perfluorinated carbons. Fluorotelomer residues denoted a possess 4, b 6, c 8 and d 10 perfluorinated carbon atoms. The four photoreactive surface modifieres 1a–d from this work are shown in Fig. 1. The fluorotelomer alcohols 3a–d employed are depicted in Fig. 2 and are readily commercially available. In these alcohols, the hydroxyl group is separated by two CH2 -units from the perfluorinated moiety for reasons of stability (a hydroxyl group directly bound to a perfluorinated carbon is not stable and decays into HF and a carbonyl compound [15]). Because photoreactive surface modifiers are to be synthesized in large quantities, their preparation has to be simple, efficient, clean and to proceed in high yields. In particular, column chromatography has to be avoided because it is unsuitable for large-scale preparations. Also, educts and reactants should be cheap, and atom economy should be high [16]. These considerations in mind, a synthesis for surface modifiers 1a–d has been elaborated (see Fig. 2). The key feature of this synthesis is to first form the ester and only then introduce the azide moiety, rather than using expensive 4-azidotetrafluorobenzoic acid as an intermediate, as it would probably be the first thought [14].
Solid, yield 99%, recrystallized from ethanol, yield 88%, mp. 44 ◦ C, ATR-IR 2137, 1732, 1644, 1490, 1232, 1185, 1141, 990 cm−1 ; 1 H NMR (500 MHz, CDCl ) ␦ 4.67 (t, J = 6.5 Hz, 2H) 2.61 (t x t, 3 J = 6.4 Hz, 3 J(19 F, 1 H) = 18.1 Hz, 2H); 13 C NMR (126 MHz, CDCl3 ) ␦ 159.0, 146.7-139.5, 124-107, 58.2 (t, 3 J(19 F, 13 C) = 4.5 Hz), 30.5 (t, 2 J(19 F, 13 C) = 22.0 Hz); 19 F NMR (471 MHz, CDCl ) ␦ −80.9 (m), 3 −113.7 (m), −121.9 (m), −122.9 (m), −123.6 (m), −126.2 (m), −138.3 (m), −150.8 (m). Anal. Calcd for C15 H4 F17 N3 O2 : C, 31.00; H, 0.69; N, 7.23. Found: C, 30.60; H, 0.58; N, 7.23. 2.11. 1H, 1H, 2H, 2H-Heptadecafluoro-1-decyl 4-Azidotetrafluorobenzoate (1c) Solid, yield 97%, recrystallized from ethanol, yield 82%, mp. 72 ◦ C, ATR-IR 2138, 1733, 1644, 1490, 1198, 1145, 991 cm−1 ; 1 H NMR (500 MHz, CDCl3 ) ␦ 4.67 (t, J = 6J = 6.5 Hz, 2H) 2.63 (t x t, J = 6.4 Hz, 3 J(19 F, 1 H) = 18.0 Hz, 2H); 13 C NMR (126 MHz, CDCl3 ) ␦ 158.9, 147–136, 119–107, 58.1 (t, 3 J(19 F, 13 C) = 4.5 Hz), 30.4 (t, 2 J(19 F, 13 C) = 22.0 Hz); 19 F NMR (471 MHz, CDCl ) ␦ −80.9 (m), 3 −113.7 (m), −121.7 (m), −122.0 (m), −122.8 (m), −123.7 (m), −126.2 (m), −138.3 (m), −150.8 (m). Anal. Calcd for C17 H4 F21 N3 O2 : C, 29.97; H, 0.59; F, 58.57; N, 6.17. Found: C, 29.76; H, 0.54; F, 58.26; N, 6.12.
O
F
+ HO
RF
F F 2
3a - d
NEt 3
RF = -C10F 21 1d
2.12. 1H, 1H, 2H, 2-Henicosafluoro-1-dodecyl 4-Azidotetrafluorobenzoate (1d)
2.10. 1H, 1H, 2H, 2H-Tridecafluoro-1-octyl 4-Azidotetrafluorobenzoate (1b)
F
F
Fig. 1. The four target molecules 1a–d.
Liquid, yield 99%, mp. 20 ◦ C, ATR-IR 2127, 1740, 1646, 1488, 1220, 1196, 1131, 1004 cm−1 ; 1 H NMR (500 MHz, CDCl3 ) ␦ 4.68 (t, J = 6.5 Hz, 2H) 2.62 (t x t, J = 6.4 Hz, 3 J(19 F, 1 H) = 18.1 Hz, 2H); 13 C NMR (126 MHz, CDCl ) ␦ 158.9, 146.6–139.4, 124–106, 58.1 3 (t, 3 J(19 F, 13 C) = 4.5 Hz), 30.3 (t, 2 J(19 F, 13 C) = 22.0 Hz); 19 F NMR (471 MHz, CDCl3 ) ␦ −81.3 (m), −114.1 (m), −124.7 (m), −126.2 (m), −138.5 (m), −151.0 (m); MALDI-TOF-MS m/z: [M + Na]+ Calcd for C13 H4 F13 N3 NaO2 , 503.9988; Found 503.9985.
F
RF = -C6F 13 1b RF = -C8F 17 1c
N3
2.9. 1H, 1H, 2H, 2H-Nonafluoro-1-hexyl 4-Azidotetrafluorobenzoate (1a)
Cl
RF = -C4F 9 1a
RF
F
F
Because azides are light sensitive they should be handled under subdued light. To a solution of 1 equiv. perfluorobenzoate 4a–d in acetone (about 0.5 molar for 4a–c; about 0.06 molar for 4d) was added 1.1 equiv. sodium azide in water (about 1.6 molar for 4a–c; about 0.2 molar for 4d). The solution was refluxed for 8 h. Thereafter, the acetone was distilled off under reduced pressure. Dichloromethane and an equal amount of water were added to the remaining. After shaking, the organic layer was separated and the water phase was extracted once with dichloromethane. The combined organic phases were dried with anhydrous sodium sulfate, filtered, and the dichloromethane was distilled off. The solid residue was recrystallized from hot ethanol.
O
O
O
2.8. General Method for the Synthesis of 4-Azidotetrafluorobenzoates (1a–d)
3
O F
F F
O
O
RF NaN3
F F 4a - d
F
F F
RF
F N3 1a - d
Fig. 2. Synthesis of photo reactive surface modifiers 1a–d.
Please cite this article in press as: K. Siegmann, et al., Photografting of perfluoroalkanes onto polyethylene surfaces via azide/nitrene chemistry, Appl. Surf. Sci. (2016), http://dx.doi.org/10.1016/j.apsusc.2016.11.007
G Model APSUSC-34309; No. of Pages 9
ARTICLE IN PRESS K. Siegmann et al. / Applied Surface Science xxx (2016) xxx–xxx
4
100
0.14
90
0.12
70
2.00*10-7 M
0.1
60
9.52*10-7 M
50
absorbance
melng point [°C]
80
40 30 20 10 0 2
4 6 8 10 number of perfluorinated carbons in the chain
0.08
1.64*10-6 M
0.06
2.26*10-6 M
3.36*10-6 M 0.02
12
Steady-state photopysical studies have been carried out for compounds 1a–d in both dissolved and neat state. The light absorbing part is the same in all compounds; they only differ in the length of the perfluorinated side chain. Hence, it is expected that the UV-spectra in solution do not differ much. Fig. 4 shows the UVabsorption spectra for compound 1c at different concentrations in acetonitrile. Indeed, the maximum of absorption is similar for all compounds 1a–d and lies at 266 nm, a typical value for benzene-based chromophores [18,19]. However, although the maxima of absorption are comparable, the molar extinction coefficients are not. Fig. 5 depicts the determination of the extinction coefficient for compound 1c. A linear fit describes the data well; the correlation coefficient (R2 ) is 0.9998. The slope of the regression line corresponds to the extinction coefficient according to the LambertBeer-model and is: 2.12• 104 ± 150 M−1 cm−1
4.29*10-6 M 200
250
300
350
400
wavelength [nm] Fig. 4. Absorption spectra of 1c in acetonitrile at different concentrations.
0.12
absorbance
0.1 0.08 0.06 0.04 0.02 0 0.E+00
1.E-06
2.E-06
3.E-06
4.E-06
5.E-06
concentraon [M] Fig. 5. Determination of molar extinction coefficient for 1c.
2.3
exncon coefficient*104 [M-1cm-1]
3.2. UV-Spectra
3.84*10-6 M
0
Fig. 3. The melting points of compounds 1a–d as a function of their fluorotelomer chain length.
Esterification of perfluorobenzoyl chloride (2) with the fluorotelomer alcohols 3a–d smoothly yields intermediates 4a–d, analogous to published procedures [17]. Yields are essentially quantitative and purification has shown to be unnecessary. The introduction of the azide moiety follows the protocol from Keana and Cai [13]. The nucleophilic aromatic substitution reaction is highly stereo selective, yielding predominantly the para-isomer, as judged by 19 F NMR. Raw yields are in the upper nineties and purified yields are high, too, although some material is lost in the recrystallization of compounds 1b–d (because of its low melting point, compound 1a could not be recrystallized). Overall, synthesis of surface modifiers 1a–d according to Fig. 2 is very efficient and can readily be carried out on a 25-gramm-scale using conventional laboratory equipment. The spectra are in accordance with the postulated structures. Fig. 3 shows the melting points of compounds 1a–d as a function of the length of the fluorotelomer chain. A linear fit is also displayed. The correlation coefficient (R2 ) of the linear fit is 0.995. It is, for example, seen that compound 1a is a liquid at room temperature, whereas compounds 1b–d are solids. The azides 1a–d were stored in a refrigerator at 4 ◦ C in the dark. The IR-spectrum of compound 1c showed no changes after about 1 year’s storage. In particular, the band at 2138 cm−1 (the stretching frequency of the azide moiety) remained unchanged. The stability of compound 1c is therefore sufficient for its intended uses. Because compounds 1a, 1b and 1d are related to compound 1c, it is anticipated that they all possess a sufficient shelf life.
2.83*10-6 M
0.04
2.2 2.1 2 1.9 1.8 1.7 1.6 1.5 1.4 2
4
6
8
10
12
number of perfluorinated carbons in the chain Fig. 6. Molar extinction coefficients, , for compounds 1a–d as a function of fluorotelomer chain length.
The extinction coefficients for compounds 1a, 1b and 1d have been determined in the same manner. Fig. 6 shows the extinction coefficients of compounds 1a–d as a function of the length of the fluorotelomer side chain.
Please cite this article in press as: K. Siegmann, et al., Photografting of perfluoroalkanes onto polyethylene surfaces via azide/nitrene chemistry, Appl. Surf. Sci. (2016), http://dx.doi.org/10.1016/j.apsusc.2016.11.007
G Model APSUSC-34309; No. of Pages 9
ARTICLE IN PRESS K. Siegmann et al. / Applied Surface Science xxx (2016) xxx–xxx
3.3. Coating and activation Fig. 8 schematically displays the intended binding of compounds 1aKd to a polyethylene surface. Presumably, the molecule becomes excited upon the absorption of a photon, followed by splitting off dinitrogen. The singlet nitrene thus produced inserts into a C H bond of the polyethylene substrate. Hence, the whole molecule is covalently—and therefore firmly—attached to the polyethylene surface. Such a binding is difficult to prove experimentally. Highresolution X-ray photoelectron spectroscopy (XPS) of the nitrogen species has been attempted in a similar case [22], and the results indicate covalent attachment but warrant further investigations. However, covalent and therefore firm binding can also be tested by abrasion and/or solubility examinations (see Section 3.5). In order for the compounds 1a–d to bind to the surface, the surface has to be clean, e.g. grease-free and dry, and a homoge-
normalized absorbance
1 0.8 1a in soluon 1a neat
0.6 0.4 0.2 0 200
250
300
350
400
wavelength [nm]
normalized absorbance
1 0.8 0.6
1b in soluon
0.4
1b neat
0.2 0 200
250
300
350
400
wavelength [nm]
normalized absorbance
1 0.8 1c in soluon 1c neat
0.6 0.4 0.2 0 200
250
300
350
400
wavelength [nm] 1
normalized absorbance
It is seen that the longer the fluorotelomer side chain, the larger is the extinction coefficient of the compound. A longer chain leads to a larger molecule, and perhaps a larger molecule possesses a higher absorption cross-section. Fluorescence measurements have been attempted. However, no sensible fluorescence (at the solvent level) could be detected for compounds 1a–d. This suggests alternative deactivation pathways for the excited states, such as UV activated chemical reactions, a very efficient vibrational relaxation, photoinduced charge separation or energy transfer. Unfortunately, the study of the excited state dynamics requires time-resolved instrumentation that is not available in our laboratories. Spectra in the neat state were obtained by spraying a solution of compounds 1a–d on a quartz plate (see Section 3.3), drying, and placing the quartz plate at the entrance of a 150 mm integrating sphere. Fig. 7a–d show the absorption spectra of compounds 1a–d measured in acetonitrile solution and on a quartz plate. The absorption spectra of 1a in solution and in the neat state are very similar, except for a low-intensity band at around 335 nm in solution. (As compound 1a was not recrystallized, this band could stem from an impurity). On the other hand, the absorption spectra obtained with compounds 1b–d on a quartz plate differ considerably from those measured in solution. The solution spectra of 1b–d show a single absorption band centered at 266 nm, whereas in the neat state they exhibit two absorption bands, a more intense one at shorter wavelengths and a less intense one at longer wavelengths. The appearance of the band at longer wavelengths in the neat state for compounds 1b–d can be explained by an exciton formation in the solid state [20]. An exciton can be found in organic systems composed of two or more similar molecules, i.e. dimers or aggregates, respectively, where the excitation energy is delocalized over the system. Exciton formation can affect the excited state of the molecules that compose the system. That is, the excited state splits into two exciton states which can be populated upon excitation, resulting into two absorption bands. The magnitude of the splitting depends on the magnitude of the interaction [21]. Assuming this scenario, one can notice that the exciton is formed in compounds 1b–d only. Yet, the exciton interaction differs within the three compounds, because the two absorption bands for each individual compound exhibit different displacement relative to each other. On the other hand, compound 1a does not form an exciton when measured in the neat state. Compound 1a also possesses a low melting point, so that it is a liquid at room temperature. Thus, we conclude that while compounds 1b–d crystallize and form excitons, compound 1a remains liquid and therefore its neat spectrum resembles the one in solution.
5
0.8 0.6
1d in soluon
0.4
1d neat
0.2 0 200
250
300
350
400
wavelength [nm] Fig. 7. a–d Absorption spectra of 1a–d in acetonitrile and on a quartz plate.
neous coating of the precursors 1a–d on the substrate has to be insured. The homogeneous coating can be achieved by dissolving compounds 1a–d in an appropriate solvent at the optimal concentration, and subsequent spray-coating. Many solvents and concentrations were tested, and it was found that butyl acetate and toluene are suited. However, by the use of perfluorinated polyether (Galden HT80) and ethoxy-nonafluorobutane (Novec HFE 7200) as solvents the most homogeneous coating was achieved, as seen by XPS. At first, concentrations of 1% and especially 0.5% of compounds 1a–d in the solvent were employed. However, it was found that a concentration as low as 0.1% yields superior results. (The relative abrasion resistances from Section 3.5 were obtained with 0.1% con-
Please cite this article in press as: K. Siegmann, et al., Photografting of perfluoroalkanes onto polyethylene surfaces via azide/nitrene chemistry, Appl. Surf. Sci. (2016), http://dx.doi.org/10.1016/j.apsusc.2016.11.007
G Model APSUSC-34309; No. of Pages 9
ARTICLE IN PRESS K. Siegmann et al. / Applied Surface Science xxx (2016) xxx–xxx
6
contact angle [degree]
120
110
100
90 Fig. 8. Binding of compounds 1a–d to a polyethylene surface.
polyethylene 1.0
70 0.8
60 50
0.6
40 0.4
30 20
0.2
10 0
80
normalized absorbance
power [mW cm-2 nm-1]
80
0.0 200 220 240 260 280 300 320 340 360 380 400 420 440
wavelength [nm] Fig. 9. A spectrum of the ozone-free mercury vapor lamp (left scale, solid line) used superimposed to an absorption spectrum of 1c in the solid state (right scale, dotted line).
centrations of 1a–d in the corresponding solvent for spray-coating and are higher compared to 0.5% concentration). Most findings in this work were obtained with 0.1% concentration, although some results were achieved with 0.5% concentration. After spray-coating the substrates were dried to insure complete removal of the solvent. Thereafter, the samples are ready for UVirradiation. An UV-belt drier equipped with a 200 W/cm low pressure mercury vapor lamp is employed for the binding process illustrated in Fig. 8. A spectrum of the ozone-free mercury vapor lamp, as it is routinely used with the belt drier, together with an absorption spectrum of compound 1c in the neat state are depicted in Fig. 9. Typical bands stemming from the mercury are observed in the emission spectrum. The glass of the light tube is doped in such a way that hard UV light below 245 nm is blocked out. Thus, ozone formation is suppressed. The absorption of 1c in the solid state overlaps with the emission of the lamp in the range of about 320–245 nm. The soft UV and visible light also emitted by the lamp does presumably therefore not participate in the nitrene formation process. With the emergence of LEDs having a narrow emission in the appropriate UV-region, photografting using azide/nitrene chemistry could therefore be made more efficient. The polyethylene samples coated with compounds 1a–d were transported below the lamp by the use of a belt. Thus, they were irradiated for a certain time depending on the speed of the belt. The standard irradiation procedure can be found in the Materials and Methods section. 3.4. Water contact angles and photoelectron spectroscopy The hydrophobicity of a surface can be determined by measuring water contact angles, hence, a contact angle as high as possible
0
10 20 30 40 surface fluorine concentraon [atom %]
50
Fig. 10. Correlation of water contact angles with fluorine surface concentration as determined by XPS for photochemically bound 1c.
is intended. Another surface-specific analysis tool is X-ray photoelectron spectroscopy (XPS). Because fluorine is sensitively probed by XPS, XPS is a well suited tool for the determination of the surface concentration of this element. Fig. 10 shows the correlation of surface fluorine content as determined by XPS with water contact angles for 1c photografted on polyethylene samples. The samples were produced as described in Section 3.3, followed by treatment with solvent or abraded with ice as described later. Two sections can be distinguished in Fig. 10: A region with linear correlation (correlation coefficient (R2 ) = 0.94) at fluorine contents below about 30%, and a region were contact angles are independent of fluorine content. Presumably, contact angles of hydrophobic surfaces cannot exceed a value of about 120◦ . As contact angles converge to this limit fluorine concentrations become negligible. We want to point out that there is yet no theory available for the linear correlation between fluorine content and water contact angle at lower surface concentrations. At present, it is just an observation. Because the values of the contact angles strongly depend on the method employed (e.g. drop size and speed of drop deposition), fluorine surface concentration measurements using XPS are used preferentially in the following sections. The slope of the regression line (Fig. 10), (contact angle)/ (fluorine concentration), is a measure for the ability of the surface modifier to increase water contact angle and is determined for surfaces photografted with compounds 1a–d. The larger this slope, the more efficient is the compound in enhancing the contact angle. Fig. 11 shows those slopes for compounds 1a–d photografted to polyethylene. It is seen, that compound 1c is most efficient when photografted to polyethylene in enhancing water contact angles. From this point of view, compound 1c would be best suited for the use as photoreactive surface modifier. However, there are other aspects as, for example, abrasion resistance, which is discussed in the next section. 3.5. Solubility—and abrasion—tests As mentioned in Section 3.3, covalent binding of compounds 1a–d, photografted to polyethylene, is difficult to prove with the spectroscopy available. We therefore set up two experiments which should demonstrate tight or permanent binding of photografted compounds 1a–d to polyethylene. The first experiment is based on the extraction of the samples with solvent in a soxhlet reactor. Thereby, the photografted polyethylene samples are con-
Please cite this article in press as: K. Siegmann, et al., Photografting of perfluoroalkanes onto polyethylene surfaces via azide/nitrene chemistry, Appl. Surf. Sci. (2016), http://dx.doi.org/10.1016/j.apsusc.2016.11.007
G Model
ARTICLE IN PRESS
APSUSC-34309; No. of Pages 9
fluorine surface content [atom %]
K. Siegmann et al. / Applied Surface Science xxx (2016) xxx–xxx
1.4 1.17
slope [degree/atom %]
1.2 1.03 1
0.884 0.8
0.723
0.6
60 50 40 30 20 10 0 0
0.4
100
200 300 me [min.]
400
Fig. 13. Fluorine surface concentration as a function of ice-induced abrasion time for polyethylene photografted with 1c (empty symbols, solid line) and a perfluoroalkane (filled symbols, dotted line).
0.2
0 compound 1a compound 1b compound 1c compound 1d Fig. 11. slope = (contact angle)/ (fluorine concentration) for surfaces photografted with compounds 1a–d.
140 118 120
108
108
104
100 80 contact angle [degree]
60 45.6
fluorine content [atom %]
40 23.3 16.8
20
13.4
0 0
7
20
40
60
80
me [h] Fig. 12. Water contact angles (dotted line) and fluorine content (solid line) as a function of time in the soxhlet apparatus for polyethylene grafted with 1c (Standard deviations are too small to be seen).
tinuously washed with freshly distilled, warm solvent. Acetone was used as solvent, because it was found that it dissolves compounds 1a–d and their exposed products well. Fig. 12 shows water contact angles and fluorine contents of the surface of a sample photografted with 1c as a function of time in the soxhlet apparatus. It is seen, that surface concentration, as determined by contact angle and XPS measurements, decays significantly after the first 24 h in the soxhlet reactor. Thereafter, surface concentration drops only slowly. However, surface concentration of photografted 1c does not reach a steady state but continuously decays, albeit slowly. Yet, if compound 1c is sprayed on a polyethylene plate and extracted in the soxhlet without prior exposition to UV light, fluorine surface concentration becomes naught after 5 h washing. This demonstrates that photografting is effective. Then why does the surface concentration of photografted1c decay with time in the soxhlet? Presumably, the solvent swells and weakens the uppermost layers of the polyethylene [23]. Thus, photografted 1c is washed from the sample, together with its anchoring hydrocarbon chain. As the tests with solvents are not unambiguous, we designed a second test that should demonstrate the stability of the pho-
tografted surfaces against abrasive wear. We compare abrasion resistance of the photografted surfaces with surfaces covered with a simple, long-chain perfluoroalkane. Hence, circular polyethylene discs (diameter = 40 mm) with a small hole in the center are photografted with compounds 1a–d. The disc is connected using a screw to the shaft of an electro-motor. This motor spins the disc in an ice-water suspension in such a way that the speed of the disc reaches 100 km/h (27.8 m/s) at its rim. The small (roughly about 5 × 5 × 5 mm in size) ice cubes collide with the photografted surface and thus abrade material from the interface. Fig. 13 shows fluorine content as measured by XPS as a function of abrasion time. Two data sets are displayed: One for polyethylene photgrafted with 1c; the other for polyethylene coated with a simple, long-chain perfluoroalkane. Exponential decay of fluorine surface concentration is observed in both cases. Exponentials are fitted for both data sets. The correlation coefficients (R2 ) are 0.994 for photografted 1c and 0.995 for the perfluoroalkane, indicating a good correlation. The exponential behavior of fluorine surface concentration versus abrasion time can be explained if one assumes a randomcollision/abrasion model. Thereby, it is assumed that the ice cubes collide with and abrade material from the surface. Because each collision of an ice cube, depending on the strength of the collision, leads to an erosion of the surface, fluorine surface concentration is reduced. The reduction of fluorine surface concentration, S, with time is proportional to the available, fluorine containing surface, S, because collisions are at random (see Eq. (1)) (In the beginning there is ample surface covered with fluorine available and almost every strong enough hit reduces fluorine surface concentration. As abrasion proceeds, fluorine covered surface becomes scarcer, there are many strong hits that miss this surface and therefore dS/dt decreases). −
dS = kS dt
Eq. (1) The reduction of fluorine containing surface (S) with time is proportional to the available fluorine containing surface. Integration of Eq. (1) yields fluorine surface concentration as a function of time (Eq. (2)). S (t) = S0 e−kt Eq. (2) Fluorine surface concentration S as a function of time. (S0 is the initial fluorine surface concentration and k is the decay constant). The exponential fit for the data points from Fig. 13 yields initial fluorine surface concentration S0 and decay constant k for surfaces photografted with 1c, as well as for surfaces covered with the perfluoroalkane. It is seen that S0 is higher for the perflu-
Please cite this article in press as: K. Siegmann, et al., Photografting of perfluoroalkanes onto polyethylene surfaces via azide/nitrene chemistry, Appl. Surf. Sci. (2016), http://dx.doi.org/10.1016/j.apsusc.2016.11.007
G Model APSUSC-34309; No. of Pages 9 8
ARTICLE IN PRESS K. Siegmann et al. / Applied Surface Science xxx (2016) xxx–xxx
surface modifier. Not all of the azides 1a–d, when photochemically transformed into nitrenes, might bind to the surface. They could produce other, unidentified reaction products (debris). This “debris” could also contribute to the relative abrasion resistance. Larger surface modifiers lead to larger and therefore stronger bound (e.g. by van der Waals’ forces) debris. Hence, the observed trend could be explained. Unfortunately, as there is little known on how the surface looks like on a molecular level, no meaningful further statement seems possible. 4. Conclusion
Fig. 14. Relative abrasion resistances for photografted surface modifiers 1a–d.
oroalkane than for surfaces photografted with 1c. (S0 = 51% for the perfluoroalkane; and S0 = 44% for photografted 1c) However, after a certain time (about an hour in Fig. 13), surface fluorine concentration S becomes larger in surfaces photografted with 1c compared to the surface covered with the perfluoroalkane. This demonstrates the enhanced durability of a surface photografted with 1c compared to a surface coated with a long-chain perfluoroalkane. Hence, |k| is larger for the perfluoroalkane than for the photoreactive surface modifier (|k| = 3.91 10−3 min−1 for the perfluoroalkane; and |k| = 1.84 10−3 min−1 for photoreactive surface modifier 1c). This observation can be explained in the randomcollision/abrasion model assuming stronger bound photoreactive surface modifier: A harder and therefore less frequent collision of an ice cube is required to remove photografted 1c compared to the perfluoroalkane. Similar behavior is observed for all compounds 1a–d, photografted to polyethylene. However, decay constants k for photografted compounds 1a–d differ markedly. In order to compare the decay constants k for photografted 1a–d and k for the perfluoroalkane among themselves, the quotient q = k(perfluoroalkane)/k(photografted surface modifier) is formed. This quotient is referred to as “relative abrasion resistance” and is plotted in Fig. 14 for compounds 1a–d and the perfluoroalkane. The relative abrasion resistance for the perfluoroalkane therefore by definition equals 1, and a number q, larger than 1, indicates a q-fold higher abrasion resistance compared to the perfluoroalkane. Because relative abrasion resistances, q, for photografted 1a and 1b were measured only once, no standard deviations are available for these q. It is seen that all compounds 1a–d, when photografted to polyethylene, display higher relative abrasion resistances than surfaces coated with the long-chain perfluoroalkane. Also, the longer the perfluorinated side chain in compounds 1a–d, the higher is the appropriate relative abrasion resistance. The reasons for the increase in abrasion resistance with increasing side chain length are not evident, because the binding, as depicted in Fig. 8, should be independent on the side chain. A possible explanation for the trend seen in Fig. 14 could be that what we observe is not only the abrasion of photografted
Four new fluorinated compounds are synthesized and characterized that should serve as photo reactive surface modifier when photografted to polyethylene. Photo physical studies in the neat state reveal that some of the molecules crystallize and form excitons when sprayed onto a quartz plate. Photografting of the four compounds to polyethylene is achieved with a lowpressure mercury vapor lamp. Water contact angle measurements and determination of the surface’s fluorine concentration by X-ray photoelectron spectroscopy are performed on grafted polyethylene and it is found that they correlate linearly for low coverage. Solubility and abrasion tests are carried out on photografted polyethylene. The washing with warm acetone yields ambiguous results, that is, the surface concentration of photo grafted surface modifier decreases with time and does not reach a steady state. Abrasion tests demonstrate superior abrasion resistance of all photo grafted surface modifiers when compared to a surface coated with a longchain perfluoroalkane. An abrasion model is introduced. It is found that the larger the grafted surface modifier in size, the higher is the relative abrasion resistance. Acknowledgement We wish to express our gratitude to Andreas Amrein and Robert Sterchi for their help with the UV-spectra. Financial support from the Commission for Technology and Innovation (KTI/CTI), Switzerland, is gratefully acknowledged. References [1] K. Kato, E. Uchida, E.T. Kang, Y. Uyama, Y. Ikada, Polymer surface with graft chains, Prog. Polym. Sci. 28 (2003) 209–259. [2] J.P. Deng, L.F. Wang, L.Y. Liu, W.T. Yang, Developments and new applications of UV-induced surface graft polymerizations, Prog. Polym. Sci. 34 (2009) 156–193. [3] S. Knaus, A. Nennadal, B. Froschauer, Surface and bulk modification of polyolefins by functional aryl nitrenes as highly reactive intermediates, Macromol. Symp. 176 (2001) 223–232. [4] V. Chowdhry, F.H. Westheimer, Photoaffinity labeling of biological systems, Annu. Rev. Biochem. 48 (1979) 293–325. [5] D.S. Novopashina, R.N. Kulikov, M.A. Kuznetsova, A.G. Venyaminova, M.A. Zenkova, V.V. Vlassov, Perfluoroarylazide derivatives of 2 ’-O-modified oligoribonucleotides: efficient reagents for RNA photomodification, Nucleosides Nucleotides Nucleic Acids 24 (2005) 1–14. [6] O. Sterner, A. Serrano, S. Mieszkin, S. Zurcher, S. Tosatti, M.E. Callow, J.A. Callow, N.D. Spencer, Photochemically prepared, two-component polymer-concentration gradients, Langmuir 29 (2013) 13031–13041. [7] V.H. Thom, G. Altankov, T. Groth, K. Jankova, G. Jonsson, M. Ulbricht, Optimizing cell-surface interactions by photografting of poly(ethylene glycol), Langmuir 16 (2000) 2756–2765. [8] V. Thom, K. Jankova, M. Ulbricht, J. Kops, G. Jonsson, Synthesis of photoreactive alpha-4-azidobenzoyl-omega-methoxypoly(ethyleneglycol)s and their end-on photo-grafting onto polysulfone ultrafiltration membranes, Macromol. Chem. Phys. 199 (1998) 2723–2729. [9] C. Wentrup, Chemical activation in azide and nitrene chemistry: methyl azide, phenyl azide, naphthyl azides, pyridyl azides, benzotriazoles, and triazolopyridines, Aust. J. Chem. 66 (2013) 852–863. [10] S. Brase, C. Gil, K. Knepper, V. Zimmermann, Organic azides: an exploding diversity of a unique class of compounds, Angew. Chem.-Int. Ed. 44 (2005) 5188–5240.
Please cite this article in press as: K. Siegmann, et al., Photografting of perfluoroalkanes onto polyethylene surfaces via azide/nitrene chemistry, Appl. Surf. Sci. (2016), http://dx.doi.org/10.1016/j.apsusc.2016.11.007
G Model APSUSC-34309; No. of Pages 9
ARTICLE IN PRESS K. Siegmann et al. / Applied Surface Science xxx (2016) xxx–xxx
[11] C. Wentrup, H. Bornemann, The Curtius rearrangement of acyl azides revisited—formation of cyanate (R-O-CN), Eur. J. Org. Chem. (2005) 4521–4524. [12] J. Brunner, New photolabeling and cross-Linking methods, Annu. Rev. Biochem. 62 (1993) 483–514. [13] J.F.W. Keana, S.X. Cai, New reagents for photoaffinity-labeling—synthesis and photolysis of functionalized perfluorophenyl azides, J. Org. Chem. 55 (1990) 3640–3647. [14] L.H. Liu, M.D. Yan, Perfluorophenyl azides: new applications in surface functionalization and nanomaterial synthesis, Acc. Chem. Res. 43 (2010) 1434–1443. [15] M.T. Nguyen, M.H. Matus, V.T. Ngan, R. Haiges, K.O. Christe, D.A. Dixon, Energetics and mechanism of the decomposition of trifluoromethanol, J. Phys. Chem. A 112 (2008) 1298–1312. [16] B.M. Trost, The atom economy—a search for synthetic efficiency, Science 254 (1991) 1471–1477.
9
[17] R. Xu, V. Gramlich, H. Frauenrath, Alternating diacetylene copolymer utilizing perfluorophenyl-phenyl interactions, J. Am. Chem. Soc. 128 (2006) 5541–5547. [18] J.B. Birks, Photophysics of Aromatic Molecules, Wiley-Interscience, 1970. [19] I.B. Berlman, Handbook of Fluorescence Spectra of Aromatic Molecules, Academic Press, 1965. [20] M. Kasha, H.R. Rawls, M.A. El-Bayoumi, The exciton model in molecular spectroscopy, Pure Appl. Chem. 11 (1965) 371–392. [21] W.W. Parson, Modern Optical Spectroscopy, Springer, 2007. [22] G. Zorn, L.H. Liu, L. Arnadottir, H. Wang, L.J. Gamble, D.G. Castner, M.D. Yan, X-ray photoelectron spectroscopy investigation of the nitrogen species in photoactive perfluorophenylazide-modified surfaces, J. Phys. Chem. C 118 (2014) 376–383. [23] K. Siegmann, A. Mosch, A. Carlesso, M. Winkler, Abrasive wear of selected plastics in solvent/corundum suspensions, Wear 324 (2015) 118–121.
Please cite this article in press as: K. Siegmann, et al., Photografting of perfluoroalkanes onto polyethylene surfaces via azide/nitrene chemistry, Appl. Surf. Sci. (2016), http://dx.doi.org/10.1016/j.apsusc.2016.11.007
ARTICLE IN PRESS
APSUSC-34309; No. of Pages 9
Applied Surface Science xxx (2016) xxx–xxx
Contents lists available at ScienceDirect
Applied Surface Science journal homepage: www.elsevier.com/locate/apsusc
Full length article
Photografting of perfluoroalkanes onto polyethylene surfaces via azide/nitrene chemistry Konstantin Siegmann a , Jan Inauen a , Diego Villamaina b , Martin Winkler a,∗ a Institute of Materials and Process Engineering (IMPE), School of Engineering (SoE), Zurich University of Applied Sciences (ZHAW), Technikumstrasse 9, CH-8401 Winterthur, Switzerland b Visiting Scientist at IMPE, Permanent Address: Rapidplatz 3, CH-8953 Dietikon, Switzerland
a r t i c l e
i n f o
Article history: Received 15 July 2016 Received in revised form 31 October 2016 Accepted 1 November 2016 Available online xxx Keywords: Nitrene Azide Insertion Polyethylene Photografting Perfluorinated compounds
a b s t r a c t Water, oil and dirt repellent surfaces of plastic materials are desirable for many applications. Polyethylene is a common plastic and, because of its inertness, difficult to graft. We chose polyethylene as example because of its ubiquity and model character. As graft chains perfluoroalkanes were chosen, and photografting was selected as grafting method. Photolytically generated nitrenes can insert into carbon-hydrogen bonds and are therefore suited for binding to polyethylene. Hydrophobic photo reactive surface modifiers based on azide/nitrene chemistry are designed, synthesized in high yield and characterized. Four new molecules are described. One problem is to demonstrate that the photografted surface modifiers are bound covalently to the polyethylene. Abrasion tests show that all new molecules, when photografted to polyethylene, have a higher abrasion resistance than a polyethylene surface coated with a long-chain perfluoroalkane. An abrasion model using ice is developed. Also, other issues are addressed, such as differences in the UVspectra of the new compounds in solution and in the neat state. Exitonic coupling of the chromophores of the surface modifiers is observed for specific molecules. A linear correlation of water contact angle with fluorine surface content, as measured by photoelectron spectroscopy, in grafted polyethylene surfaces is established. © 2016 Published by Elsevier B.V.
1. Introduction Highly hydrophobic surfaces are desirable for many purposes, such as water, oil and dirt repellent outfit of polymeric materials. Polyethylene finds widespread uses as commodity plastic in a wealth of consumer goods. We therefore set ourselves to modify polyethylene surfaces in such a way that they become strongly hydrophobic. Further, this hydrophobicity should be resistant against abrasion and wear. From the point of view of a chemist, polyethylene is interesting as an example polymer, because it consists of carbon–carbon and carbon-hydrogen bonds only, which make it very inert and difficult to modify. If polyethylene surfaces can be rendered strongly and permanently hydrophobic, then other plastics could presumably be treated in the same way. Perfluorinated alky chains are known to possess a high water- and
∗ Corresponding author. E-mail addresses: [email protected] (K. Siegmann), [email protected] (J. Inauen), [email protected] (D. Villamaina), [email protected] (M. Winkler).
oil repellency which makes them apt to be used as hydrophobic surface modifiers. The task therefore is to graft polyethylene with perfluorinated alky chains. The modification of polymer surfaces has become an important theme for many applications [1]. From the methods developed, surface grafting has emerged as simple, useful and versatile approach. One of the major techniques is ultra violet (UV) light induced grafting, in all of its variations [2]. Yet unreactive surfaces, such as the surfaces of polyolefines, can be grafted using UV-light [3]. However, drastic conditions have to be used which were originally developed for photoaffinity labeling in bio-chemistry [4]. This method relies on photo chemically generated singlet nitrenes. A singlet nitrene is a nitrogen species having only six valence electrons (a sextet), that is highly reactive and therefore capable of inserting even into carbon-hydrogen bonds of the substrate [5,6]. This approach has been applied to surfaces, too. Two publications describe the surface modification of polysulfon with polyethylene glycol using photo chemically generated nitrenes for the prevention of cell adhesion [7,8]. The precursor molecule for the generation of a nitrene species usually is an azide [9,10]. Organic azides intermediately form
http://dx.doi.org/10.1016/j.apsusc.2016.11.007 0169-4332/© 2016 Published by Elsevier B.V.
Please cite this article in press as: K. Siegmann, et al., Photografting of perfluoroalkanes onto polyethylene surfaces via azide/nitrene chemistry, Appl. Surf. Sci. (2016), http://dx.doi.org/10.1016/j.apsusc.2016.11.007
G Model APSUSC-34309; No. of Pages 9
ARTICLE IN PRESS K. Siegmann et al. / Applied Surface Science xxx (2016) xxx–xxx
2
nitrenes by the uptake of energy, such as temperature or light, followed by the elimination of dinitrogen. However, because nitrenes are highly reactive, rearrangements of the parent molecule are frequent [11]. In order to convert azides into nitrenes by UV-radiation, the azide has to be bound to a chromophore (a light absorbing part), that is, an aromatic moiety. Since aromatic systems are stabilized by resonance, rearrangements of the intermediate nitrene are somewhat suppressed. However, one of the main reactions of phenyl nitrenes is their ring expansion to azacycloheptatetraenes which react further and do not yield the insertion product [12]. Hence, perfluorinated phenyl nitrenes were investigated and it was found that by their use rearrangements can be hindered in such a way that the desired carbon-hydrogen insertion reaction becomes dominant [12,13]. Adhesion promotors on the basis of photoreactive perfluorophenyl azide derivatives have been developed [14]. The desired hydrophobic and oleophobic effect is caused by the highly water repellent perfluorinated carbon chain. Such a chain has to be attached to the perfluorophenyl azide in order to obtain a photo reactive surface modifier. As fluorotelomer alcohols are readily available (they are extensively employed as water and dirt repellant substituents in the textile industry) it immediately suggests itself to use those alcohols as the hydrophobic part in the new molecules. Linking the fluorotelomer alcohol to the perfluorophenyl azide moiety is conveniently achieved by an ester functionality. The target molecule is therefore composed of four segments: The azide as the reactive group, the perfluorinated aromatic ring as chromophore and stabilizator, the ester functionality as linker and the fluorotelomer alcohol as the hydrophobic part. In the following sections synthesis, photochemistry and stability studies of the new molecules and surfaces are described.
2. Materials and methods 2.1. Materials All Chemicals were used without further purification. Fluorocarbon chemicals were purchased from Apollo Scientific Ltd, UK. Anhydrous dichloromethane was purchased from TCI Deutschland GmbH. Sodium azide (purum, p. a., ≥99.0%) and triethylamine (puriss, p. a., ≥99.5%) were from Sigma-Aldrich Chemie GmbH, Switzerland. Other chemicals were purchased from Sigma-Aldrich Chemie GmbH, Switzerland. Polyethylene was a high-density Borstar ME3440 type from Borealis, and the polyethylene plates had a thickness of 2 mm.
2.2. Instrumentation Melting points were determined on a Differential Scanning Calorimeter, DSC 204 F1 Phoenix from Netzsch Gerätebau GmbH. UV-Spectra were recorded on a Perkin Elmer 950 Spectrophotometer, equipped with a 150 mm integrating sphere for the neat spectra; fluorescence spectra on a LS 55 Fluorescence Spectrometer from Perkin Elmer. ATR-IR spectra were recorded using a Perkin Elmer Spectrum 100 FT-IR Spectrometer. Most NMR Spectra were run on a Bruker AscendTM 500 instrument. The MALDI-MS and the elemental analyses were from the Lab. for Organic Chemistry, ETH Zurich, Switzerland. Static water contact angles were measured using a Krüss DSA 100 instrument. XPS Spectra were obtained using a SPECSTM spectrometer from SPECS GmbH, Berlin, Germany. The UV radiometer was an UVpad spectral radiometer from Opsytec Dr. Gröbel GmbH, Germany. The UV belt drier was from Uviterno AG, Switzerland, equipped with a 150 mm light tube with a maximal power of 3 kW. Standard irradiation procedure was as follows: Speed of the belt: 0.026 m/s; lamp power 1.2 kW; dis-
tance lamp-sample 0.07 m. The spray-coated polyethylene samples were exposed 2 to 3 times to these conditions. 2.3. General method for the synthesis of pentafluorobenzoates (4a–d) The reaction was kept under an inert atmosphere of dry nitrogen at all times. To a stirred solution of 1 equiv. fluorotelomer alcohol 3a–d in dry dichloromethane (about 0.3 molar for alcohols 3a–c; about 0.05 molar for alcohol 3d) at 0 ◦ C was added 1.05 equiv. perfluorobenzoyl chloride (2) via a syringe. Then, 2.1 equiv. of triethylamine were added via a syringe. The mixture was stirred for 2 h at 0 ◦ C, allowed to warm to room temperature and stirred overnight. Work up: The solution was washed twice with equal amounts of 1 molar hydrochloric acid. Then, the organic phase was washed once with saturated sodium bicarbonate solution and once with water. The organic phase was dried with anhydrous sodium sulfate, filtered, and the dichloromethane was distilled off. The perfluorobenzoates were not purified but used as obtained for the next step. 2.4. 1H, 1H, 2H, 2H-Nonafluoro-1-hexyl Pentafluorobenzoate (4a) Liquid, yield 95% ATR-IR 1746, 1653, 1499, 1331, 1216, 1132, 1007 cm−1 ; 1 H NMR (500 MHz, CDCl3 ) ␦ 4.70 (t, J = 6.4 Hz, 2H) 2.63 (t x t, J = 6.4 Hz, 3 J(19 F, 1 H) = 18.1 Hz, 2H); 13 C NMR (126 MHz, CDCl3 ) ␦ 158.6, 146.7–136.6, 121–107, 58.3 t, 3 J(19 F, 13 C) = 4.6 Hz, 30.3 (t, 2 J(19 F, 13 C) = 22.0 Hz); 19 F NMR (471 MHz, CDCl3 ) ␦ −81.4 (m), −114.1 (m), −124.7 (m), −126.2 (m), −138.1 (m), −148 (m), −160.5 (m). 2.5. 1H, 1H, 2H, 2H-Tridecafluoro-1-octyl Pentafluorobenzoate (4b) Liquid, yield 96% ATR-IR 1746, 1653, 1500, 1331, 1190, 1144, 1009 cm−1 ; 1 H NMR (500 MHz, CDCl3 ) ␦ 4.70 (t, J = 6.5 Hz, 2H) 2.63 (t x t, J = 6.4 Hz, 3 J(19 F, 1 H) = 18.1 Hz, 2H); 13 C NMR (126 MHz, CDCl3 ) ␦ 158.6, 146.7–136.6, 121–107, 58.2 (t, 3 J(19 F, 13 C) = 4.5 Hz), 30.3 (t, 2 J(19 F, 13 C) = 21.8 Hz); 19 F NMR (471 MHz, CDCl3 ) ␦ −81.5 (m), −114.1 (m), −122.3 (m), −123.3 (m) −126.2 (m), −124.0 (m), −126.6 (m), −138.4 (m), −148.5 (m), −161.0 (m). 2.6. 1H, 1H, 2H, 2H-Heptadecafluoro-1-decyl Pentafluorobenzoate (4c) Solid, yield 97% ATR-IR 1738, 1652, 1501, 1332, 1197, 1145, 987 cm−1 ; 1 H NMR (300 MHz, CDCl3 ) ␦ 4.69 (t, J = 6.4 Hz, 2H) 2.61 (t x t, J = 6.4 Hz, 3 J(19 F, 1 H) = 18.0 Hz, 2H); 13 C NMR (75 MHz, CDCl3 ) ␦ 158.6, 147–136, 119–107, 58.3 (t, 3 J(19 F, 13 C) = 4.8 Hz), 30.4 (t, 2 J(19 F, 13 C) = 22.0 Hz); 19 F NMR (282 MHz, CDCl ) ␦ −81.1 (m), 3 −113.8 (m), −121.8 (m), −122.1 (m), −122.9 (m), −123.7 (m), −126.3 (m), −138.0 (m), −147.9 (m), −160.4 (m). 2.7. 1H, 1H, 2H, 2H Henicosafluoro-1-dodecyl Pentafluorobenzoate (4d) Solid, yield 97% ATR-IR 1738, 1653, 1529, 1497, 1199, 1147, 1006 cm−1 ; 1 H NMR (500 MHz, CDCl3 ) ␦ 4.68 (t, J = 6.4 Hz, 2H) 2.61 (t x t, J = 6.4 Hz, 3 J(19 F, 1 H) = 18.1 Hz, 2H); 13 C NMR (126 MHz, CDCl3 ) ␦ 158.8, 147–137, 119–107, 58.5 (t, 3 J(19 F, 13 C) = 4.6 Hz), 30.6 (t, 2 J(19 F, 13 C) = 21.9 Hz); 19 F NMR (471 MHz, CDCl3 ) ␦ −81.0 (m), −113.8 (m), −121.9 (m), −122.0 (m), −122.8 (m), −123.7 (m), −126.3 (m), −137.9 (m), −147.7 (m), −160.3 (m).
Please cite this article in press as: K. Siegmann, et al., Photografting of perfluoroalkanes onto polyethylene surfaces via azide/nitrene chemistry, Appl. Surf. Sci. (2016), http://dx.doi.org/10.1016/j.apsusc.2016.11.007
G Model
ARTICLE IN PRESS
APSUSC-34309; No. of Pages 9
K. Siegmann et al. / Applied Surface Science xxx (2016) xxx–xxx
F
Solid, yield 95%, recrystallized from ethanol, yield 74%, mp. 91 ◦ C, ATR-IR 2139, 1733, 1644, 1491, 1198, 1148, 997 cm−1 ; 1 H NMR (500 MHz, CDCl3 ) ␦ 4.67 (t, J = 6.4 Hz, 2H) 2.61 (t x t, J = 6.5 Hz, 3 J(19 F, 1 H) = 18.0 Hz, 2H); 13 C NMR (126 MHz, CDCl ) ␦ 159.0, 3 146.7–139.5, 124–107, 58.2 (t, 3 J(19 F, 13 C) = 4.5 Hz), 30.5 (t, 2 J(19 F, 13 C) = 21.7 Hz); 19 F NMR (471 MHz, CDCl ) ␦ −80.8 (m), −113.7 3 (m), −121.8 (m), −122.8 (m), −123.6 (m), −126.2 (m), −138.3 (m), −150.8 (m). Anal. Calcd for C19 H4 F25 N3 O2 : C, 29.21; H, 0.52; N, 5.38. Found: C, 29.12; H, 0.49; N, 5.52.
3. Results and discussion 3.1. Synthesis As described in the Introduction section, a photoreactive surface modifier consists of four segments. These can be seen in Fig. 1: the azide as reactive group, the perfluororophenyl moiety as chromophore, the ester functionality as linker and a fluorotelomer alcohol as the hydrophobic part. Four different fluorotelomer residues, RF , were investigated, and they are assigned to the letters a, b, c, and d throughout the paper. The fluorotelomer residues are non-branched, linear chains with an even number of perfluorinated carbons. Fluorotelomer residues denoted a possess 4, b 6, c 8 and d 10 perfluorinated carbon atoms. The four photoreactive surface modifieres 1a–d from this work are shown in Fig. 1. The fluorotelomer alcohols 3a–d employed are depicted in Fig. 2 and are readily commercially available. In these alcohols, the hydroxyl group is separated by two CH2 -units from the perfluorinated moiety for reasons of stability (a hydroxyl group directly bound to a perfluorinated carbon is not stable and decays into HF and a carbonyl compound [15]). Because photoreactive surface modifiers are to be synthesized in large quantities, their preparation has to be simple, efficient, clean and to proceed in high yields. In particular, column chromatography has to be avoided because it is unsuitable for large-scale preparations. Also, educts and reactants should be cheap, and atom economy should be high [16]. These considerations in mind, a synthesis for surface modifiers 1a–d has been elaborated (see Fig. 2). The key feature of this synthesis is to first form the ester and only then introduce the azide moiety, rather than using expensive 4-azidotetrafluorobenzoic acid as an intermediate, as it would probably be the first thought [14].
Solid, yield 99%, recrystallized from ethanol, yield 88%, mp. 44 ◦ C, ATR-IR 2137, 1732, 1644, 1490, 1232, 1185, 1141, 990 cm−1 ; 1 H NMR (500 MHz, CDCl ) ␦ 4.67 (t, J = 6.5 Hz, 2H) 2.61 (t x t, 3 J = 6.4 Hz, 3 J(19 F, 1 H) = 18.1 Hz, 2H); 13 C NMR (126 MHz, CDCl3 ) ␦ 159.0, 146.7-139.5, 124-107, 58.2 (t, 3 J(19 F, 13 C) = 4.5 Hz), 30.5 (t, 2 J(19 F, 13 C) = 22.0 Hz); 19 F NMR (471 MHz, CDCl ) ␦ −80.9 (m), 3 −113.7 (m), −121.9 (m), −122.9 (m), −123.6 (m), −126.2 (m), −138.3 (m), −150.8 (m). Anal. Calcd for C15 H4 F17 N3 O2 : C, 31.00; H, 0.69; N, 7.23. Found: C, 30.60; H, 0.58; N, 7.23. 2.11. 1H, 1H, 2H, 2H-Heptadecafluoro-1-decyl 4-Azidotetrafluorobenzoate (1c) Solid, yield 97%, recrystallized from ethanol, yield 82%, mp. 72 ◦ C, ATR-IR 2138, 1733, 1644, 1490, 1198, 1145, 991 cm−1 ; 1 H NMR (500 MHz, CDCl3 ) ␦ 4.67 (t, J = 6J = 6.5 Hz, 2H) 2.63 (t x t, J = 6.4 Hz, 3 J(19 F, 1 H) = 18.0 Hz, 2H); 13 C NMR (126 MHz, CDCl3 ) ␦ 158.9, 147–136, 119–107, 58.1 (t, 3 J(19 F, 13 C) = 4.5 Hz), 30.4 (t, 2 J(19 F, 13 C) = 22.0 Hz); 19 F NMR (471 MHz, CDCl ) ␦ −80.9 (m), 3 −113.7 (m), −121.7 (m), −122.0 (m), −122.8 (m), −123.7 (m), −126.2 (m), −138.3 (m), −150.8 (m). Anal. Calcd for C17 H4 F21 N3 O2 : C, 29.97; H, 0.59; F, 58.57; N, 6.17. Found: C, 29.76; H, 0.54; F, 58.26; N, 6.12.
O
F
+ HO
RF
F F 2
3a - d
NEt 3
RF = -C10F 21 1d
2.12. 1H, 1H, 2H, 2-Henicosafluoro-1-dodecyl 4-Azidotetrafluorobenzoate (1d)
2.10. 1H, 1H, 2H, 2H-Tridecafluoro-1-octyl 4-Azidotetrafluorobenzoate (1b)
F
F
Fig. 1. The four target molecules 1a–d.
Liquid, yield 99%, mp. 20 ◦ C, ATR-IR 2127, 1740, 1646, 1488, 1220, 1196, 1131, 1004 cm−1 ; 1 H NMR (500 MHz, CDCl3 ) ␦ 4.68 (t, J = 6.5 Hz, 2H) 2.62 (t x t, J = 6.4 Hz, 3 J(19 F, 1 H) = 18.1 Hz, 2H); 13 C NMR (126 MHz, CDCl ) ␦ 158.9, 146.6–139.4, 124–106, 58.1 3 (t, 3 J(19 F, 13 C) = 4.5 Hz), 30.3 (t, 2 J(19 F, 13 C) = 22.0 Hz); 19 F NMR (471 MHz, CDCl3 ) ␦ −81.3 (m), −114.1 (m), −124.7 (m), −126.2 (m), −138.5 (m), −151.0 (m); MALDI-TOF-MS m/z: [M + Na]+ Calcd for C13 H4 F13 N3 NaO2 , 503.9988; Found 503.9985.
F
RF = -C6F 13 1b RF = -C8F 17 1c
N3
2.9. 1H, 1H, 2H, 2H-Nonafluoro-1-hexyl 4-Azidotetrafluorobenzoate (1a)
Cl
RF = -C4F 9 1a
RF
F
F
Because azides are light sensitive they should be handled under subdued light. To a solution of 1 equiv. perfluorobenzoate 4a–d in acetone (about 0.5 molar for 4a–c; about 0.06 molar for 4d) was added 1.1 equiv. sodium azide in water (about 1.6 molar for 4a–c; about 0.2 molar for 4d). The solution was refluxed for 8 h. Thereafter, the acetone was distilled off under reduced pressure. Dichloromethane and an equal amount of water were added to the remaining. After shaking, the organic layer was separated and the water phase was extracted once with dichloromethane. The combined organic phases were dried with anhydrous sodium sulfate, filtered, and the dichloromethane was distilled off. The solid residue was recrystallized from hot ethanol.
O
O
O
2.8. General Method for the Synthesis of 4-Azidotetrafluorobenzoates (1a–d)
3
O F
F F
O
O
RF NaN3
F F 4a - d
F
F F
RF
F N3 1a - d
Fig. 2. Synthesis of photo reactive surface modifiers 1a–d.
Please cite this article in press as: K. Siegmann, et al., Photografting of perfluoroalkanes onto polyethylene surfaces via azide/nitrene chemistry, Appl. Surf. Sci. (2016), http://dx.doi.org/10.1016/j.apsusc.2016.11.007
G Model APSUSC-34309; No. of Pages 9
ARTICLE IN PRESS K. Siegmann et al. / Applied Surface Science xxx (2016) xxx–xxx
4
100
0.14
90
0.12
70
2.00*10-7 M
0.1
60
9.52*10-7 M
50
absorbance
melng point [°C]
80
40 30 20 10 0 2
4 6 8 10 number of perfluorinated carbons in the chain
0.08
1.64*10-6 M
0.06
2.26*10-6 M
3.36*10-6 M 0.02
12
Steady-state photopysical studies have been carried out for compounds 1a–d in both dissolved and neat state. The light absorbing part is the same in all compounds; they only differ in the length of the perfluorinated side chain. Hence, it is expected that the UV-spectra in solution do not differ much. Fig. 4 shows the UVabsorption spectra for compound 1c at different concentrations in acetonitrile. Indeed, the maximum of absorption is similar for all compounds 1a–d and lies at 266 nm, a typical value for benzene-based chromophores [18,19]. However, although the maxima of absorption are comparable, the molar extinction coefficients are not. Fig. 5 depicts the determination of the extinction coefficient for compound 1c. A linear fit describes the data well; the correlation coefficient (R2 ) is 0.9998. The slope of the regression line corresponds to the extinction coefficient according to the LambertBeer-model and is: 2.12• 104 ± 150 M−1 cm−1
4.29*10-6 M 200
250
300
350
400
wavelength [nm] Fig. 4. Absorption spectra of 1c in acetonitrile at different concentrations.
0.12
absorbance
0.1 0.08 0.06 0.04 0.02 0 0.E+00
1.E-06
2.E-06
3.E-06
4.E-06
5.E-06
concentraon [M] Fig. 5. Determination of molar extinction coefficient for 1c.
2.3
exncon coefficient*104 [M-1cm-1]
3.2. UV-Spectra
3.84*10-6 M
0
Fig. 3. The melting points of compounds 1a–d as a function of their fluorotelomer chain length.
Esterification of perfluorobenzoyl chloride (2) with the fluorotelomer alcohols 3a–d smoothly yields intermediates 4a–d, analogous to published procedures [17]. Yields are essentially quantitative and purification has shown to be unnecessary. The introduction of the azide moiety follows the protocol from Keana and Cai [13]. The nucleophilic aromatic substitution reaction is highly stereo selective, yielding predominantly the para-isomer, as judged by 19 F NMR. Raw yields are in the upper nineties and purified yields are high, too, although some material is lost in the recrystallization of compounds 1b–d (because of its low melting point, compound 1a could not be recrystallized). Overall, synthesis of surface modifiers 1a–d according to Fig. 2 is very efficient and can readily be carried out on a 25-gramm-scale using conventional laboratory equipment. The spectra are in accordance with the postulated structures. Fig. 3 shows the melting points of compounds 1a–d as a function of the length of the fluorotelomer chain. A linear fit is also displayed. The correlation coefficient (R2 ) of the linear fit is 0.995. It is, for example, seen that compound 1a is a liquid at room temperature, whereas compounds 1b–d are solids. The azides 1a–d were stored in a refrigerator at 4 ◦ C in the dark. The IR-spectrum of compound 1c showed no changes after about 1 year’s storage. In particular, the band at 2138 cm−1 (the stretching frequency of the azide moiety) remained unchanged. The stability of compound 1c is therefore sufficient for its intended uses. Because compounds 1a, 1b and 1d are related to compound 1c, it is anticipated that they all possess a sufficient shelf life.
2.83*10-6 M
0.04
2.2 2.1 2 1.9 1.8 1.7 1.6 1.5 1.4 2
4
6
8
10
12
number of perfluorinated carbons in the chain Fig. 6. Molar extinction coefficients, , for compounds 1a–d as a function of fluorotelomer chain length.
The extinction coefficients for compounds 1a, 1b and 1d have been determined in the same manner. Fig. 6 shows the extinction coefficients of compounds 1a–d as a function of the length of the fluorotelomer side chain.
Please cite this article in press as: K. Siegmann, et al., Photografting of perfluoroalkanes onto polyethylene surfaces via azide/nitrene chemistry, Appl. Surf. Sci. (2016), http://dx.doi.org/10.1016/j.apsusc.2016.11.007
G Model APSUSC-34309; No. of Pages 9
ARTICLE IN PRESS K. Siegmann et al. / Applied Surface Science xxx (2016) xxx–xxx
3.3. Coating and activation Fig. 8 schematically displays the intended binding of compounds 1aKd to a polyethylene surface. Presumably, the molecule becomes excited upon the absorption of a photon, followed by splitting off dinitrogen. The singlet nitrene thus produced inserts into a C H bond of the polyethylene substrate. Hence, the whole molecule is covalently—and therefore firmly—attached to the polyethylene surface. Such a binding is difficult to prove experimentally. Highresolution X-ray photoelectron spectroscopy (XPS) of the nitrogen species has been attempted in a similar case [22], and the results indicate covalent attachment but warrant further investigations. However, covalent and therefore firm binding can also be tested by abrasion and/or solubility examinations (see Section 3.5). In order for the compounds 1a–d to bind to the surface, the surface has to be clean, e.g. grease-free and dry, and a homoge-
normalized absorbance
1 0.8 1a in soluon 1a neat
0.6 0.4 0.2 0 200
250
300
350
400
wavelength [nm]
normalized absorbance
1 0.8 0.6
1b in soluon
0.4
1b neat
0.2 0 200
250
300
350
400
wavelength [nm]
normalized absorbance
1 0.8 1c in soluon 1c neat
0.6 0.4 0.2 0 200
250
300
350
400
wavelength [nm] 1
normalized absorbance
It is seen that the longer the fluorotelomer side chain, the larger is the extinction coefficient of the compound. A longer chain leads to a larger molecule, and perhaps a larger molecule possesses a higher absorption cross-section. Fluorescence measurements have been attempted. However, no sensible fluorescence (at the solvent level) could be detected for compounds 1a–d. This suggests alternative deactivation pathways for the excited states, such as UV activated chemical reactions, a very efficient vibrational relaxation, photoinduced charge separation or energy transfer. Unfortunately, the study of the excited state dynamics requires time-resolved instrumentation that is not available in our laboratories. Spectra in the neat state were obtained by spraying a solution of compounds 1a–d on a quartz plate (see Section 3.3), drying, and placing the quartz plate at the entrance of a 150 mm integrating sphere. Fig. 7a–d show the absorption spectra of compounds 1a–d measured in acetonitrile solution and on a quartz plate. The absorption spectra of 1a in solution and in the neat state are very similar, except for a low-intensity band at around 335 nm in solution. (As compound 1a was not recrystallized, this band could stem from an impurity). On the other hand, the absorption spectra obtained with compounds 1b–d on a quartz plate differ considerably from those measured in solution. The solution spectra of 1b–d show a single absorption band centered at 266 nm, whereas in the neat state they exhibit two absorption bands, a more intense one at shorter wavelengths and a less intense one at longer wavelengths. The appearance of the band at longer wavelengths in the neat state for compounds 1b–d can be explained by an exciton formation in the solid state [20]. An exciton can be found in organic systems composed of two or more similar molecules, i.e. dimers or aggregates, respectively, where the excitation energy is delocalized over the system. Exciton formation can affect the excited state of the molecules that compose the system. That is, the excited state splits into two exciton states which can be populated upon excitation, resulting into two absorption bands. The magnitude of the splitting depends on the magnitude of the interaction [21]. Assuming this scenario, one can notice that the exciton is formed in compounds 1b–d only. Yet, the exciton interaction differs within the three compounds, because the two absorption bands for each individual compound exhibit different displacement relative to each other. On the other hand, compound 1a does not form an exciton when measured in the neat state. Compound 1a also possesses a low melting point, so that it is a liquid at room temperature. Thus, we conclude that while compounds 1b–d crystallize and form excitons, compound 1a remains liquid and therefore its neat spectrum resembles the one in solution.
5
0.8 0.6
1d in soluon
0.4
1d neat
0.2 0 200
250
300
350
400
wavelength [nm] Fig. 7. a–d Absorption spectra of 1a–d in acetonitrile and on a quartz plate.
neous coating of the precursors 1a–d on the substrate has to be insured. The homogeneous coating can be achieved by dissolving compounds 1a–d in an appropriate solvent at the optimal concentration, and subsequent spray-coating. Many solvents and concentrations were tested, and it was found that butyl acetate and toluene are suited. However, by the use of perfluorinated polyether (Galden HT80) and ethoxy-nonafluorobutane (Novec HFE 7200) as solvents the most homogeneous coating was achieved, as seen by XPS. At first, concentrations of 1% and especially 0.5% of compounds 1a–d in the solvent were employed. However, it was found that a concentration as low as 0.1% yields superior results. (The relative abrasion resistances from Section 3.5 were obtained with 0.1% con-
Please cite this article in press as: K. Siegmann, et al., Photografting of perfluoroalkanes onto polyethylene surfaces via azide/nitrene chemistry, Appl. Surf. Sci. (2016), http://dx.doi.org/10.1016/j.apsusc.2016.11.007
G Model APSUSC-34309; No. of Pages 9
ARTICLE IN PRESS K. Siegmann et al. / Applied Surface Science xxx (2016) xxx–xxx
6
contact angle [degree]
120
110
100
90 Fig. 8. Binding of compounds 1a–d to a polyethylene surface.
polyethylene 1.0
70 0.8
60 50
0.6
40 0.4
30 20
0.2
10 0
80
normalized absorbance
power [mW cm-2 nm-1]
80
0.0 200 220 240 260 280 300 320 340 360 380 400 420 440
wavelength [nm] Fig. 9. A spectrum of the ozone-free mercury vapor lamp (left scale, solid line) used superimposed to an absorption spectrum of 1c in the solid state (right scale, dotted line).
centrations of 1a–d in the corresponding solvent for spray-coating and are higher compared to 0.5% concentration). Most findings in this work were obtained with 0.1% concentration, although some results were achieved with 0.5% concentration. After spray-coating the substrates were dried to insure complete removal of the solvent. Thereafter, the samples are ready for UVirradiation. An UV-belt drier equipped with a 200 W/cm low pressure mercury vapor lamp is employed for the binding process illustrated in Fig. 8. A spectrum of the ozone-free mercury vapor lamp, as it is routinely used with the belt drier, together with an absorption spectrum of compound 1c in the neat state are depicted in Fig. 9. Typical bands stemming from the mercury are observed in the emission spectrum. The glass of the light tube is doped in such a way that hard UV light below 245 nm is blocked out. Thus, ozone formation is suppressed. The absorption of 1c in the solid state overlaps with the emission of the lamp in the range of about 320–245 nm. The soft UV and visible light also emitted by the lamp does presumably therefore not participate in the nitrene formation process. With the emergence of LEDs having a narrow emission in the appropriate UV-region, photografting using azide/nitrene chemistry could therefore be made more efficient. The polyethylene samples coated with compounds 1a–d were transported below the lamp by the use of a belt. Thus, they were irradiated for a certain time depending on the speed of the belt. The standard irradiation procedure can be found in the Materials and Methods section. 3.4. Water contact angles and photoelectron spectroscopy The hydrophobicity of a surface can be determined by measuring water contact angles, hence, a contact angle as high as possible
0
10 20 30 40 surface fluorine concentraon [atom %]
50
Fig. 10. Correlation of water contact angles with fluorine surface concentration as determined by XPS for photochemically bound 1c.
is intended. Another surface-specific analysis tool is X-ray photoelectron spectroscopy (XPS). Because fluorine is sensitively probed by XPS, XPS is a well suited tool for the determination of the surface concentration of this element. Fig. 10 shows the correlation of surface fluorine content as determined by XPS with water contact angles for 1c photografted on polyethylene samples. The samples were produced as described in Section 3.3, followed by treatment with solvent or abraded with ice as described later. Two sections can be distinguished in Fig. 10: A region with linear correlation (correlation coefficient (R2 ) = 0.94) at fluorine contents below about 30%, and a region were contact angles are independent of fluorine content. Presumably, contact angles of hydrophobic surfaces cannot exceed a value of about 120◦ . As contact angles converge to this limit fluorine concentrations become negligible. We want to point out that there is yet no theory available for the linear correlation between fluorine content and water contact angle at lower surface concentrations. At present, it is just an observation. Because the values of the contact angles strongly depend on the method employed (e.g. drop size and speed of drop deposition), fluorine surface concentration measurements using XPS are used preferentially in the following sections. The slope of the regression line (Fig. 10), (contact angle)/ (fluorine concentration), is a measure for the ability of the surface modifier to increase water contact angle and is determined for surfaces photografted with compounds 1a–d. The larger this slope, the more efficient is the compound in enhancing the contact angle. Fig. 11 shows those slopes for compounds 1a–d photografted to polyethylene. It is seen, that compound 1c is most efficient when photografted to polyethylene in enhancing water contact angles. From this point of view, compound 1c would be best suited for the use as photoreactive surface modifier. However, there are other aspects as, for example, abrasion resistance, which is discussed in the next section. 3.5. Solubility—and abrasion—tests As mentioned in Section 3.3, covalent binding of compounds 1a–d, photografted to polyethylene, is difficult to prove with the spectroscopy available. We therefore set up two experiments which should demonstrate tight or permanent binding of photografted compounds 1a–d to polyethylene. The first experiment is based on the extraction of the samples with solvent in a soxhlet reactor. Thereby, the photografted polyethylene samples are con-
Please cite this article in press as: K. Siegmann, et al., Photografting of perfluoroalkanes onto polyethylene surfaces via azide/nitrene chemistry, Appl. Surf. Sci. (2016), http://dx.doi.org/10.1016/j.apsusc.2016.11.007
G Model
ARTICLE IN PRESS
APSUSC-34309; No. of Pages 9
fluorine surface content [atom %]
K. Siegmann et al. / Applied Surface Science xxx (2016) xxx–xxx
1.4 1.17
slope [degree/atom %]
1.2 1.03 1
0.884 0.8
0.723
0.6
60 50 40 30 20 10 0 0
0.4
100
200 300 me [min.]
400
Fig. 13. Fluorine surface concentration as a function of ice-induced abrasion time for polyethylene photografted with 1c (empty symbols, solid line) and a perfluoroalkane (filled symbols, dotted line).
0.2
0 compound 1a compound 1b compound 1c compound 1d Fig. 11. slope = (contact angle)/ (fluorine concentration) for surfaces photografted with compounds 1a–d.
140 118 120
108
108
104
100 80 contact angle [degree]
60 45.6
fluorine content [atom %]
40 23.3 16.8
20
13.4
0 0
7
20
40
60
80
me [h] Fig. 12. Water contact angles (dotted line) and fluorine content (solid line) as a function of time in the soxhlet apparatus for polyethylene grafted with 1c (Standard deviations are too small to be seen).
tinuously washed with freshly distilled, warm solvent. Acetone was used as solvent, because it was found that it dissolves compounds 1a–d and their exposed products well. Fig. 12 shows water contact angles and fluorine contents of the surface of a sample photografted with 1c as a function of time in the soxhlet apparatus. It is seen, that surface concentration, as determined by contact angle and XPS measurements, decays significantly after the first 24 h in the soxhlet reactor. Thereafter, surface concentration drops only slowly. However, surface concentration of photografted 1c does not reach a steady state but continuously decays, albeit slowly. Yet, if compound 1c is sprayed on a polyethylene plate and extracted in the soxhlet without prior exposition to UV light, fluorine surface concentration becomes naught after 5 h washing. This demonstrates that photografting is effective. Then why does the surface concentration of photografted1c decay with time in the soxhlet? Presumably, the solvent swells and weakens the uppermost layers of the polyethylene [23]. Thus, photografted 1c is washed from the sample, together with its anchoring hydrocarbon chain. As the tests with solvents are not unambiguous, we designed a second test that should demonstrate the stability of the pho-
tografted surfaces against abrasive wear. We compare abrasion resistance of the photografted surfaces with surfaces covered with a simple, long-chain perfluoroalkane. Hence, circular polyethylene discs (diameter = 40 mm) with a small hole in the center are photografted with compounds 1a–d. The disc is connected using a screw to the shaft of an electro-motor. This motor spins the disc in an ice-water suspension in such a way that the speed of the disc reaches 100 km/h (27.8 m/s) at its rim. The small (roughly about 5 × 5 × 5 mm in size) ice cubes collide with the photografted surface and thus abrade material from the interface. Fig. 13 shows fluorine content as measured by XPS as a function of abrasion time. Two data sets are displayed: One for polyethylene photgrafted with 1c; the other for polyethylene coated with a simple, long-chain perfluoroalkane. Exponential decay of fluorine surface concentration is observed in both cases. Exponentials are fitted for both data sets. The correlation coefficients (R2 ) are 0.994 for photografted 1c and 0.995 for the perfluoroalkane, indicating a good correlation. The exponential behavior of fluorine surface concentration versus abrasion time can be explained if one assumes a randomcollision/abrasion model. Thereby, it is assumed that the ice cubes collide with and abrade material from the surface. Because each collision of an ice cube, depending on the strength of the collision, leads to an erosion of the surface, fluorine surface concentration is reduced. The reduction of fluorine surface concentration, S, with time is proportional to the available, fluorine containing surface, S, because collisions are at random (see Eq. (1)) (In the beginning there is ample surface covered with fluorine available and almost every strong enough hit reduces fluorine surface concentration. As abrasion proceeds, fluorine covered surface becomes scarcer, there are many strong hits that miss this surface and therefore dS/dt decreases). −
dS = kS dt
Eq. (1) The reduction of fluorine containing surface (S) with time is proportional to the available fluorine containing surface. Integration of Eq. (1) yields fluorine surface concentration as a function of time (Eq. (2)). S (t) = S0 e−kt Eq. (2) Fluorine surface concentration S as a function of time. (S0 is the initial fluorine surface concentration and k is the decay constant). The exponential fit for the data points from Fig. 13 yields initial fluorine surface concentration S0 and decay constant k for surfaces photografted with 1c, as well as for surfaces covered with the perfluoroalkane. It is seen that S0 is higher for the perflu-
Please cite this article in press as: K. Siegmann, et al., Photografting of perfluoroalkanes onto polyethylene surfaces via azide/nitrene chemistry, Appl. Surf. Sci. (2016), http://dx.doi.org/10.1016/j.apsusc.2016.11.007
G Model APSUSC-34309; No. of Pages 9 8
ARTICLE IN PRESS K. Siegmann et al. / Applied Surface Science xxx (2016) xxx–xxx
surface modifier. Not all of the azides 1a–d, when photochemically transformed into nitrenes, might bind to the surface. They could produce other, unidentified reaction products (debris). This “debris” could also contribute to the relative abrasion resistance. Larger surface modifiers lead to larger and therefore stronger bound (e.g. by van der Waals’ forces) debris. Hence, the observed trend could be explained. Unfortunately, as there is little known on how the surface looks like on a molecular level, no meaningful further statement seems possible. 4. Conclusion
Fig. 14. Relative abrasion resistances for photografted surface modifiers 1a–d.
oroalkane than for surfaces photografted with 1c. (S0 = 51% for the perfluoroalkane; and S0 = 44% for photografted 1c) However, after a certain time (about an hour in Fig. 13), surface fluorine concentration S becomes larger in surfaces photografted with 1c compared to the surface covered with the perfluoroalkane. This demonstrates the enhanced durability of a surface photografted with 1c compared to a surface coated with a long-chain perfluoroalkane. Hence, |k| is larger for the perfluoroalkane than for the photoreactive surface modifier (|k| = 3.91 10−3 min−1 for the perfluoroalkane; and |k| = 1.84 10−3 min−1 for photoreactive surface modifier 1c). This observation can be explained in the randomcollision/abrasion model assuming stronger bound photoreactive surface modifier: A harder and therefore less frequent collision of an ice cube is required to remove photografted 1c compared to the perfluoroalkane. Similar behavior is observed for all compounds 1a–d, photografted to polyethylene. However, decay constants k for photografted compounds 1a–d differ markedly. In order to compare the decay constants k for photografted 1a–d and k for the perfluoroalkane among themselves, the quotient q = k(perfluoroalkane)/k(photografted surface modifier) is formed. This quotient is referred to as “relative abrasion resistance” and is plotted in Fig. 14 for compounds 1a–d and the perfluoroalkane. The relative abrasion resistance for the perfluoroalkane therefore by definition equals 1, and a number q, larger than 1, indicates a q-fold higher abrasion resistance compared to the perfluoroalkane. Because relative abrasion resistances, q, for photografted 1a and 1b were measured only once, no standard deviations are available for these q. It is seen that all compounds 1a–d, when photografted to polyethylene, display higher relative abrasion resistances than surfaces coated with the long-chain perfluoroalkane. Also, the longer the perfluorinated side chain in compounds 1a–d, the higher is the appropriate relative abrasion resistance. The reasons for the increase in abrasion resistance with increasing side chain length are not evident, because the binding, as depicted in Fig. 8, should be independent on the side chain. A possible explanation for the trend seen in Fig. 14 could be that what we observe is not only the abrasion of photografted
Four new fluorinated compounds are synthesized and characterized that should serve as photo reactive surface modifier when photografted to polyethylene. Photo physical studies in the neat state reveal that some of the molecules crystallize and form excitons when sprayed onto a quartz plate. Photografting of the four compounds to polyethylene is achieved with a lowpressure mercury vapor lamp. Water contact angle measurements and determination of the surface’s fluorine concentration by X-ray photoelectron spectroscopy are performed on grafted polyethylene and it is found that they correlate linearly for low coverage. Solubility and abrasion tests are carried out on photografted polyethylene. The washing with warm acetone yields ambiguous results, that is, the surface concentration of photo grafted surface modifier decreases with time and does not reach a steady state. Abrasion tests demonstrate superior abrasion resistance of all photo grafted surface modifiers when compared to a surface coated with a longchain perfluoroalkane. An abrasion model is introduced. It is found that the larger the grafted surface modifier in size, the higher is the relative abrasion resistance. Acknowledgement We wish to express our gratitude to Andreas Amrein and Robert Sterchi for their help with the UV-spectra. Financial support from the Commission for Technology and Innovation (KTI/CTI), Switzerland, is gratefully acknowledged. References [1] K. Kato, E. Uchida, E.T. Kang, Y. Uyama, Y. Ikada, Polymer surface with graft chains, Prog. Polym. Sci. 28 (2003) 209–259. [2] J.P. Deng, L.F. Wang, L.Y. Liu, W.T. Yang, Developments and new applications of UV-induced surface graft polymerizations, Prog. Polym. Sci. 34 (2009) 156–193. [3] S. Knaus, A. Nennadal, B. Froschauer, Surface and bulk modification of polyolefins by functional aryl nitrenes as highly reactive intermediates, Macromol. Symp. 176 (2001) 223–232. [4] V. Chowdhry, F.H. Westheimer, Photoaffinity labeling of biological systems, Annu. Rev. Biochem. 48 (1979) 293–325. [5] D.S. Novopashina, R.N. Kulikov, M.A. Kuznetsova, A.G. Venyaminova, M.A. Zenkova, V.V. Vlassov, Perfluoroarylazide derivatives of 2 ’-O-modified oligoribonucleotides: efficient reagents for RNA photomodification, Nucleosides Nucleotides Nucleic Acids 24 (2005) 1–14. [6] O. Sterner, A. Serrano, S. Mieszkin, S. Zurcher, S. Tosatti, M.E. Callow, J.A. Callow, N.D. Spencer, Photochemically prepared, two-component polymer-concentration gradients, Langmuir 29 (2013) 13031–13041. [7] V.H. Thom, G. Altankov, T. Groth, K. Jankova, G. Jonsson, M. Ulbricht, Optimizing cell-surface interactions by photografting of poly(ethylene glycol), Langmuir 16 (2000) 2756–2765. [8] V. Thom, K. Jankova, M. Ulbricht, J. Kops, G. Jonsson, Synthesis of photoreactive alpha-4-azidobenzoyl-omega-methoxypoly(ethyleneglycol)s and their end-on photo-grafting onto polysulfone ultrafiltration membranes, Macromol. Chem. Phys. 199 (1998) 2723–2729. [9] C. Wentrup, Chemical activation in azide and nitrene chemistry: methyl azide, phenyl azide, naphthyl azides, pyridyl azides, benzotriazoles, and triazolopyridines, Aust. J. Chem. 66 (2013) 852–863. [10] S. Brase, C. Gil, K. Knepper, V. Zimmermann, Organic azides: an exploding diversity of a unique class of compounds, Angew. Chem.-Int. Ed. 44 (2005) 5188–5240.
Please cite this article in press as: K. Siegmann, et al., Photografting of perfluoroalkanes onto polyethylene surfaces via azide/nitrene chemistry, Appl. Surf. Sci. (2016), http://dx.doi.org/10.1016/j.apsusc.2016.11.007
G Model APSUSC-34309; No. of Pages 9
ARTICLE IN PRESS K. Siegmann et al. / Applied Surface Science xxx (2016) xxx–xxx
[11] C. Wentrup, H. Bornemann, The Curtius rearrangement of acyl azides revisited—formation of cyanate (R-O-CN), Eur. J. Org. Chem. (2005) 4521–4524. [12] J. Brunner, New photolabeling and cross-Linking methods, Annu. Rev. Biochem. 62 (1993) 483–514. [13] J.F.W. Keana, S.X. Cai, New reagents for photoaffinity-labeling—synthesis and photolysis of functionalized perfluorophenyl azides, J. Org. Chem. 55 (1990) 3640–3647. [14] L.H. Liu, M.D. Yan, Perfluorophenyl azides: new applications in surface functionalization and nanomaterial synthesis, Acc. Chem. Res. 43 (2010) 1434–1443. [15] M.T. Nguyen, M.H. Matus, V.T. Ngan, R. Haiges, K.O. Christe, D.A. Dixon, Energetics and mechanism of the decomposition of trifluoromethanol, J. Phys. Chem. A 112 (2008) 1298–1312. [16] B.M. Trost, The atom economy—a search for synthetic efficiency, Science 254 (1991) 1471–1477.
9
[17] R. Xu, V. Gramlich, H. Frauenrath, Alternating diacetylene copolymer utilizing perfluorophenyl-phenyl interactions, J. Am. Chem. Soc. 128 (2006) 5541–5547. [18] J.B. Birks, Photophysics of Aromatic Molecules, Wiley-Interscience, 1970. [19] I.B. Berlman, Handbook of Fluorescence Spectra of Aromatic Molecules, Academic Press, 1965. [20] M. Kasha, H.R. Rawls, M.A. El-Bayoumi, The exciton model in molecular spectroscopy, Pure Appl. Chem. 11 (1965) 371–392. [21] W.W. Parson, Modern Optical Spectroscopy, Springer, 2007. [22] G. Zorn, L.H. Liu, L. Arnadottir, H. Wang, L.J. Gamble, D.G. Castner, M.D. Yan, X-ray photoelectron spectroscopy investigation of the nitrogen species in photoactive perfluorophenylazide-modified surfaces, J. Phys. Chem. C 118 (2014) 376–383. [23] K. Siegmann, A. Mosch, A. Carlesso, M. Winkler, Abrasive wear of selected plastics in solvent/corundum suspensions, Wear 324 (2015) 118–121.
Please cite this article in press as: K. Siegmann, et al., Photografting of perfluoroalkanes onto polyethylene surfaces via azide/nitrene chemistry, Appl. Surf. Sci. (2016), http://dx.doi.org/10.1016/j.apsusc.2016.11.007