Poly[POSS-(amic acid)]s were synthesized by interfacial polymerization of octa-ammonium ..... Here, a Weibull-distribution was chosen to model the distributed ...
Preprint of Supporting Information for: Thermal Imidization Kinetics of Ultrathin Films of Hybrid Poly(POSS imide)s, Macromolecules, 48 (9), pp. 3031-3039. DOI: 10.1021/acs.macromol.5b00473.
Thermal Imidization Kinetics of Ultrathin Films of Hybrid Poly(POSS imide)s - Supporting Information Michiel J.T. Raaijmakers†,‡, Emiel J. Kappert †,‡, Arian Nijmeijer † and Nieck E. Benes *,† †
Inorganic Membranes, University of Twente, Department of Science and Technology, MESA+
Institute for Nanotechnology, P.O. Box 217, 7500 AE Enschede, The Netherlands
PolyPOSS-imide synthesis Poly[POSS-(amic acid)]s were synthesized by interfacial polymerization of octa-ammonium functionalized POSS in water and a dianhydride solution in toluene. Five different dianhydrides were used as precursor for the poly[POSS-(amic acid)]s; 4,4-(hexafluoroisopropylidene) diphthalic
anhydride
(6FDA),
pyromellitic
dianhydride
(PMDA),
3,3',4,4'-biphenyl
tetracarboxylic dianhydride (BPDA), 4,4'-oxydiphthalic anhydride (ODPA), 4,4’-(4,4’isopropylidene diphenoxy) bis(phthalic anhydride) (BPADA). Scheme S1 shows an overview of the synthesis of the poly[POSS-(amic acid)]s from the octa-ammonium POSS and dianhydride
precursors, and the subsequent imidization of the poly[POSS-(amic acid)] to a polyimide. The octa-ammonium POSS is first dissolved in water. The ammonium groups are partially deprotonated using NaOH, to form the more reactive amine groups. The solution is subsequently contacted with the dianhydride solution in toluene. A heat treatment is applied to convert the amic acid groups to cyclic imides. Scheme S1. Schematic representation of the synthesis of polyPOSS-imides via interfacial polymerization. (a)
R R Si
O
Si
Si
R'
O Si O
O
Si
Si
NaOH
O
R O Si
R
Si
RO
O O
R
R
O
O Si O
R' O
O
+ (CH2)3 NH3 Cl
O
R' O Si Si O
R'
R
R=
Si O
Si
O
-HCl
R
R'
O
O Si O
O Si O
Si
R
R'
R' =
Pyromellitic dianhydride (PMDA)
(CH2)3 NH2
O
O
O
O
3,3’,4,4’-biphenyl tetracarboxylic dianhydride (BPDA)
(c)
O
O
O
O
O
O
POSS R1
O
O
HN O
O
O
O
Si O
O O
O
O
O
O
3FC
O
CF3
O
(H2C)3 O
O
R1
O
O
O
Si
O
Si
O
O O Si O
Si
(CH2)3
HN
R'
(H2C)3 Si
(CH2)3 NH R1
O Si O
O
(H2C)3 R1 POSS
HN
(H2C)3 Si O
O
O Si O
Si
(CH2)3
POSS
R'
O H N
O
HO OH H N
O
R R'
(H2C)3
O
Si
(CH2)3
O Si O
O
Si O
Si
O O
R1 HN (H2C)3
Si
R1=amic acid R2=imide
O R'
O Si POSS
O Si O
O
O O Si O
Si
R'
(CH2)3 NH R1
Conversion processes during thermal imidization
POSS
Si O
Si
O R'
O Si
Amic acid
O
(CH2)3 O
O
O R'
O Si Si
R'
N
Imide
O
O O
O
O
POSS
Si
Si
O N
O
POSS
HN
R'
R
POSS R1
POSS
Si O
R'
O Si
O
O
4,4-(hexafluoroisopropylidene) diphthalic anhydride (6FDA)
(CH2)3 NH R1
O Si O O
O
O
4,4’-(4,4’-isopropylidinediphenoxy) bis(phthalic anhydride) (BPADA)
(b)
R'
(H2C)3
O
4,4’oxydiphthalic anhydride (ODPA)
R1 HN (H2C)3
Si
O
O O Si O
Si
R'
(CH2)3 NH R1
POSS
During the heat treatment of the poly[POSS-(amic acid)]s, reaction steps associated with drying and imidization (1+2), silanol condensation (3) and decomposition (4) are observed. Scheme S2 shows the reaction mechanism of step 2 and 3 that occur. During step 2 of the heat treatment process, mainly conversion of amic acid to imide groups occurs (blue). During step 3, additional water loss is observed that originates from recombination of silanol groups to form siloxane groups (red). Scheme S2. Overview of the water loss originating from imidization and silanol condensation reactions.
Imidization
1. amic acid O HO
-H2O 2. imide F3 C
CF3
O
H N
N O
SiOH recombination to Si-O-Si
(H2C)3
O
Si
O Si O
O
Si
Si
O OH
O
OH
HO
3. Silanol
O O Si
HO Si
O Si O
Si
O
4. Siloxane
-H2O
The silanol condensation reaction is observed from attenuated total reflection – Fourier transform infrared spectroscopy (ATR-FTIR) measurements of poly[POSS-(amic acid)] samples that were heat treated at different temperatures. Figure S1 shows the relative intensity of infrared peaks that
are characteristic for Si-O bonds in a cage or ladder (i.e., a broken POSS cage) structure, and peaks that are characteristic for silanol groups. The peak at 3230 cm-1 could include vibrations from water present in the material. Although the scatter in the data makes a quantitative analysis difficult, the data shows a trend of increasing ladder and cage formation with increasing treatment temperatures, at the expense of silanol groups.
Relative peak intensity (au.)
SiO ladder at 1040 cm-1
1.0 SiOH @ 910 cm-1
1.2 SiO cage @ 1090 cm-1
0.6
1.0
SiOH @ 3230 cm-1
0
100
200
Temperature (°C)
0.8
300
400
0
100
200
300
0.4
Relative peak intensity (au.)
1.2
1.4
0.2 400
Temperature (°C)
Figure S1. (left panel) Relative intensity of the infrared peaks at 1040 and 1090 cm-1, corresponding to a Si-O-Si bond in ladder and cage configuration, respectively; (right panel) Relative intensity of the infrared peaks at 910 and 3230 cm-1, both corresponding to silanol groups. All infrared spectra were normalized with respect to the CF3 band at 1254 cm−1. All peak intensities are normalized with respect to their initial peak intensity. Thickness and refractive indices The thickness and refractive index have been measured using spectroscopic ellipsometry. Figure S2 shows the evolution of the thickness (top panels) and refractive index (bottom panels) of polyPOSS-(amic acid)s prepared using PMDA, BPDA, ODPA, BPADA and 6FDA precursors,
as function of temperature (left panel) and as function of time during the subsequent dwell at 345 °C (right panel). The data in Figure S2 correspond to the relative thickness data in Figure 3 and relative refractive index data in Figure 4.
160
160
PMDA
140
BPDA
120
120 ODPA
100
100
80
Thickness (nm)
Thickness (nm)
140
80
BPADA 6FDA
0
50
100
150
200
250
300
0
50
100
Temperature (°C)
200
250
300
60
Time (min)
1.52
Refractive index (-)
150
1.52
ODPA
1.50
1.50 6FDA
1.48
1.48 BPADA PMDA
1.46
Refractive index (-)
60
1.46
BPDA
1.44
0
50
100
150
200
250
Temperature (°C)
300
0
50
100
150
200
250
300
1.44
Time (min)
Figure S2. (top panels) Thickness and (bottom panels) refractive index of the polyPOSS-(amic acid)s layer as function of temperature (left), and dwell time at 345 °C after heating from 50 to
345 °C at a heating rate of 5 °C min-1. The polyPOSS-(amic acid) layers were prepared atop a γalumina coated α-alumina disc using PMDA, BPDA, ODPA, BPADA and 6FDA precursors. Duplicate measurements were performed to validate the optical model of the spectroscopic ellipsometry measurements. Figure S3 shows the refractive index and thickness data of two polyPOSS-(amic acid) prepared using ODPA (left panel). The data shows a similar evolution in thickness and refractive index. The absolute values of the thickness are slightly different, which can be related to the reproducibility of the interfacial polymerization reaction.
1.525
140
1.520 n (-)
Thickness (nm)
150
130 1.515 120
110
0
100
200
300
1.510
Temperature (°C)
Figure S3. Refractive index and thickness data of two polyPOSS-(amic acid)s prepared using ODPA (left panel). The linear thermal expansion coefficient of the polyPOSS-imides has been determined by measuring the thickness at 50 °C and 345 °C.
Table S1 shows the linear thermal expansion coefficients of the polyPOSS-imides
Table S1. Linear thermal expansion coefficient of polyPOSS-imides prepared using PMDA, BPDA, ODPA, BPADA and 6FDA
Dianhydride linker PMDA BPDA ODPA BPADA 6FDA
Thickness at 50 °C (nm)
Thickness at 345 °C (nm)
Linear thermal expansion coefficient (10-6 C-1)
118.6 113.2 91.9 64.4 59.2
121.3 114.6 95.7 70.5 61.9
76 44 141 317 154
Kinetic analysis Figure S4 (left panel) shows the conversion of the mass loss of 6FDA-POSS as a function of temperature for heating rates β=5,10,15 and 20 °C min-1. The shift in the curves introduced by the different heating rates is an indication of kinetically limited processes. A plateau in the conversion curves is related to the absence of mass loss processes. The curves before and after the plateau can therefore be appointed to distinct, serial reaction steps. The conversion profile demonstrates 3 distinct mass loss steps: step 1 from α = 0-0.16, step 2 from α = 0.16-0.21, and step 3 from α = 0.21-1. The effective activation energy for mass loss as a function of the conversion can be calculated by performing an isoconversional analysis.1 This activation energy is referred to as an ‘effective’ activation energy, as it is in fact an average over the activation energies of the different reactions taking place at a specific value of the conversion. The
occurrence of multiple reactions within the distinct steps is apparent from the release of different fragments that is observed in the TGA-MS data. Step 1 is dominated by the loss of water from the material, accompanied by the release of small amounts of CO2 and aromatic fragments. Step 2 is only coupled to the loss of water. In step 3, aromatics, CO2 and water are released. Figure S4
Conversion (-)
1.0
5 °C min-1 10 °C min-1 15 °C min-1 20 °C min-1
0.5
0.0 100
200
300
400
500
600
Temperature (°C)
(right panel) shows the effective activation energy for the thermal imidization steps of poly(POSS-amic acid) prepared with 6FDA. The three distinct steps result in three plateaus in the effective activation energy: For step 1, Ea ~225 kJ mol-1 ; for step 2, Ea ~150 kJ mol-1; for step 3, Ea increased from ~225 to ~400 kJ mol-1.
Conversion (-)
1.0
5 °C min-1 10 °C min-1 15 °C min-1 20 °C min-1
0.5
0.0 100
200
300
400
500
600
Temperature (°C)
Figure S4. (left panel) Conversion of the mass loss of 6FDA-based poly[POSS-(amic acid)] as function of temperature, for heating rates β=5,10,15 and 20 °C min-1 under N2 atmosphere. (right panel) Apparent activation energy as function of conversion, determined by the isoconversional analysis of TGA-data for the thermal treatment of 6FDA based poly[POSS-(amic acid)]. Only an activation energy is insufficient for performing kinetic calculations, as the reaction rate is typically expressed as a function of A, the pre-exponential constant, and f(α). An overview of possible reaction models is given in Table S2.
Table S2. Overview of possible reaction models, taken from 2 Reaction model (name + abbreviation) Power law – P4 Power law – P3 Power law – P2 Power law – P2/3 Mampel (first order) – F1 Avrami-Erofeev – A4
f(α) 4α3/4 3α2/3 2α1/2 2/3α1/2 1-α 4(1-α)[-ln(1-α)]3/4
Avrami-Erofeev – A3 Avrami-Erofeev – A2 Contracting sphere – R3 Contracting cylinder – R2 One-dimensional diffusion – D1 Two-dimensional diffusion – D2 Three-dimensional diffusion (Jander) – D3 Three-dimensional diffusion (GinstlingBrounshtein) D4
3(1-α)[-ln(1-α)]2/3 2(1-α)[-ln(1-α)]1/2 3(1-α)2/3 2(1-α)1/2 1/2α-1 [-ln(1-α)]-1 3/2(1-α)2/3[1-(1-α)1/3]-1 3/2[(1-α)-1/3-1]-1
In order to get insight into the reaction model, a temperature program containing isothermal dwells was ran. The temperatures of the isothermals were chosen such that in every isothermal step, one of the reactions was taking place. In line with previous observations, the profile of the conversion versus the time revealed a decelerating reaction profile that is commonly associated to the reaction order (F) or diffusion (D) models from Table S2. 3-7 To determine which reaction model captures the data most adequately, and to obtain the value for the pre-exponential constant, the data given in Figure S4 (left panel) was fit by a multivariate kinetic model. To perform this analysis, the 3 steps in the activation energy were first fitted individually, using the activation energies obtained from Figure S4 right panel) as starting values. The data was fit 14 times, one time for each reaction models given in Table S2. These individual fits were followed by a combined fit of the full conversion range. For the fitting of the first step, it was found that the data could not be captured using a single kinetic triplet (A, Ea, and f(α)). This is of no surprise, as the MS-data gave clear evidence for multiple reactions occurring in this step. Surprisingly, however, fitting of the data with a combination of two individual reactions did not result in a good fit either. Therefore, it was tried to fit the data with a distributed activation energy. Here, a Weibull-distribution was chosen to model the distributed activation energy. For certain values of the parameters, this distribution can
be mathematically equivalent to a Gamma-distributed activation energy or a nth order reaction (cite: Burnham&Braun 1999 Energy & Fuels 13 1-22). Using the Weibull-distribution, the data could be accurately fit, as long as the fitting was started from a conversion value of α = 0.02. Below this value, the individual curves do not demonstrate a shift with heating rate, implying a thermodynamic equilibrium. Step 2 could be captured accurately using a single kinetic triplet. Tests for the different reaction models revealed that a first order reaction (F1) model captured the data most accurately. For step 3, again a distribution in the activation energy was required for the fitting. Again this is with no surprise, since the degradation of the organic fragments typically consist of many different reactions, smeared out over a broad temperature range. These reaction distributions are commonly fit with distributed kinetic parameters.8 Figures S5-9 show the overview of all m/z-signals of the mass spectroscopy measurements as a function of the temperature for poly[POSS-(amic acid)]s prepared using the different dianhydrides.
Figure S5. m/z-signals of the mass spectroscopy measurements as a function of the temperature for poly[POSS-(amic acid)]s prepared using 6FDA.
Figure S6. m/z-signals of the mass spectroscopy measurements as a function of the temperature for poly[POSS-(amic acid)]s prepared using BPADA.
Figure S7. m/z-signals of the mass spectroscopy measurements as a function of the temperature for poly[POSS-(amic acid)]s prepared using BPDA.
Figure S8. m/z-signals of the mass spectroscopy measurements as a function of the temperature for poly[POSS-(amic acid)]s prepared using ODPA.
Figure S9. m/z-signals of the mass spectroscopy measurements as a function of the temperature for poly[POSS-(amic acid)]s prepared using PMDA.
References 1. Vyazovkin, S.; Burnham, A. K.; Criado, J. M.; Pérez-Maqueda, L. A.; Popescu, C.; Sbirrazzuoli, N. Thermochimica Acta 2011, 520, (1-2), 1-19. 2. Vyazovkin, S.; Chrissafis, K.; Di Lorenzo, M. L.; Koga, N.; Pijolat, M.; Roduit, B.; Sbirrazzuoli, N.; Suñol, J. J. Thermochimica Acta 2014, 590, 1-23. 3. Seo, Y.; Lee, S. M.; Kim, D. Y.; Kim, K. U. Macromolecules 1997, 30, (13), 3747-3753. 4. Yilmaz, T.; Güçlü, H.; Özarslan, Ö.; Yildiz, E.; Kuyulu, A.; Ekinci, E.; Güngör, A. Journal of Polymer Science, Part A: Polymer Chemistry 1997, 35, (14), 2981-2990. 5. Seo, Y. Polymer Engineering and Science 1997, 37, (5), 772-776. 6. Kim, Y. J.; Glass, T. E.; Lyle, G. D.; McGrath, J. E. Macromolecules 1993, 26, (6), 13441358. 7. Lu, H.; Zhou, J.; He, T. Journal of Applied Polymer Science 2001, 79, (11), 2052-2059. 8. Burnham, A. K.; Braun, R. L. Energy & Fuels 1999, 13, (1), 1-22.
