Anal. Chem. 1996, 68, 3033-3037
Quantitative Analysis and Structure Determination of Styrene/Methyl Methacrylate Copolymers by Pyrolysis Gas Chromatography Frank Cheng-Yu Wang* and Patrick B. Smith
Analytical Sciences Laboratory, Michigan Division, The Dow Chemical Company, Midland, Michigan 48667
A pyrolysis gas chromatography (Py-GC) method has been developed to study the composition and microstructure of styrene/methyl methacrylate (STY/MMA) copolymers. The composition was quantified by Py-GC using monomer peak intensity. Because of the poor stability of methyl methacrylate oligomers, neither MMA dimer nor MMA trimers were detected under normal pyrolysis conditions. The number-average sequence length for STY was determined by pure and hybrid trimer peak intensities. The number-average sequence length for MMA was determined by using formulas that incorporate composition and the number-average sequence length of STY. This method is a new approach for the investigation of the microstructure of those copolymers that do not produce dimer and trimer peaks upon pyrolysis. The structural study of polymeric materials through dimers and trimers by pyrolysis gas chromatography (Py-GC) has been developed over a long time.1-5 No matter which approach has been used to investigate the structure, the basic requirement is that the dimer and trimer peaks have to be detected and identified. For polymer systems containing monomers that will not produce detectable dimer and trimer peaks such as R-methylstyrene and aliphatic methacrylates, acceptable methods for structure determinations using dimer and trimer peaks are not available. Styrene and methyl methacrylate (STY/MMA) copolymers are produced by radical polymerization using suspension or bulk processes.6 The presence of methyl methacrylate in the polymer improves the resistance to nonpolar solvents, especially aromatic hydrocarbons. The other application properties such as weatherability and hardness, have also been enhanced.7 The STY/MMA copolymers are historically important, as they were employed in early theoretical studies of problems in coploymerization and reactivity ratios.8,9 (1) Kalal, J.; Zachoval, J.; Kubat, J.; Svec, F. J. Anal. Appl. Pyrolysis 1979, 1, 143-157. (2) Shimono, T.; Tanaka, M.; Shono, T. J. Anal. Appl. Pyrolysis 1979, 1, 7784. (3) Alajberg, A.; Arpino, P.; Deur-Siftar, D.; Guiochon, G. J. Anal. Appl. Pyrolysis 1980, 1, 203-12. (4) Wang, F. C.-Y.; Gerhart, B. B.; Smith, P. B. Anal. Chem. 1995, 67, 353640. (5) Wang, F. C.-Y.; Smith, P. B. Anal. Chem. 1996, 68, 425-30. (6) Svec, P.; Rosik, L.; Horak, L. Z.; Vecerka, F. Styrene-Based Plastic and Modification; Ellis Horwood: Chichester, U.K., 1989; pp 103-4. (7) Davies, T. E. Brit. Plast. 1960, 33 (5), 195. (8) Mayo, F. R.; Lewis, F. M. J. Am. Chem. Soc. 1944, 66, 1954. (9) Pepper, D. C. Chem. Soc. Q. Rev. 1954, 8, 88. S0003-2700(96)00196-5 CCC: $12.00
© 1996 American Chemical Society
The composition of STY/MMA copolymer systems has been studied by liquid absorption chromatography.10,11 The structure of STY/MMA random copolymer systems has also been studied by combining Py-GC with a modeling method.2 In this study, PyGC has been used to produce very clear and distinguishable monomer peaks that are used in the determination of the composition. In addition, the partially available dimer and trimer peaks, also from Py-GC, are used in combination with the composition to determine the number-average sequence length and percentage of grouped monomers.4 The trimer peaks of STY/ MMA copolymers were identified by Py-GC/MS in both electron ionization (EI) and chemical ionization (CI) mode. The compositions of the copolymers were compared to the recipe amounts or compared with 13C NMR results, when available. THEORY The number-average sequence length is most commonly used as an indicator of the degree of polymer structure. The statistical method to calculate the number-average sequence length from both dimer and trimer distributions has been known for more than a decade.12 The method is widely used in NMR spectroscopic analysis because NMR (especially 13C NMR) can supply diad and triad molar fractions. In a copolymer system containing monomers S and M, the number-average sequence length of monomers can be calculated from either diad or triad molar fractions, the formulas expressed as
n ˜S )
n ˜M )
n ˜S )
n ˜M +
NSS + NMS 1
/2NMS
NMM + NMS 1
/2NMS
NSSS + NMSS+SSM + NMSM 1
/2NMSS+SSM + NMSM
NMMM + NSMM+MMS + NSMS 1
/2NSMM+MMS + NSMS
(1)
(2)
(3)
(4)
where n ˜S and n ˜M are the number-average sequence length of monomers S and M. NSS, NSM, NMM, and NSSS, NSSM+MSS, NMSM, NSMS, NSMM+MMS, and NMMM are the experimentally derived three and six distinguishable diad and triad molar fractions or number (10) Mori, S.; Uno, Y. Anal. Chem. 1986, 58, 303. (11) Mori, S.; Uno, Y. Anal. Chem. 1987, 59, 90. (12) Randall, J. C. Polymer Sequence Determination; Academic Press: New York, 1977; pp 41-69.
Analytical Chemistry, Vol. 68, No. 17, September 1, 1996 3033
of molecules. From the formula above, if all three dimer or six trimer molar fractions or number of molecules can be generated, the number-average sequence length of monomer S and M can be calculated. The number-average sequence length can be further used to calculate the molar percent of monomer in the copolymer. In a copolymer system containing monomers S and M, the molar percentage of S and M in terms of the number-average sequence length of both monomers is expressed as
mol % S )
n ˜S × 100 n ˜S + n ˜M
resulting in the series of expressions
Cstd1 k1Astd1 ) Cstd2 k2Astd2 Cstd1 k1Astd1 ) Cstd3 k3Astd3 Cstd1 k1Astd1 ) Cstd4 k4Astd4
(5) l
mol % M )
n ˜M × 100 n ˜S + n ˜M
Utilization of the monomer-type molecular weights will allow for conversion to weight percent, which is the commonly used unit when designing and synthesizing copolymers. Additionally, the weight percent of the monomers can be calculated from the monomer peak intensity, especially when monomers are the major pyrolysis products. When a copolymer system produces incomplete dimers and trimer peaks because one of the monomers does not produce detectable dimers and trimers, the structural information still can be explored if the number-average sequence length of one monomer can be calculated through dimer or trimer peak intensities and the other sequence distribution can be derived from that which is known together with the composition data. Analysis of a polymer composition by Py-GC depends on the monomer retention index for qualitative analysis. When there is a linear relationship between the concentration of monomer contained in the polymer and pyrolysis peak intensity of that monomer, the monomer peak area can be used for quantitative analysis. Because of the difficulty of controlling the exact amount of sample loaded for repeat measurements and the nature of stability of the pyrolysis process, it will be easier if the quantitative analysis of polymer composition depends on the relative peak area ratio rather than the absolute peak area. The following formula has been developed based on the relative peak area ratio for the quantitative analysis of a polymer system with n monomers. (A) Calculation of Proportionality Constants (k’s) from a standard. Given a polymer standard comprised of n monomer components of known concentration Cstd1, Cstd2, Cstd3, ..., Cstdn (in mole percent or weight percent), there is a corresponding experimentally determined peak area Astd1, Astd2, Astd3, ..., Astdn such that
Cstdn ) knAstdn
Cstd1 k1Astd1 ) Cstdn knAstdn
(6)
if we now define K12 ) k1/k2, K13 ) k1/k3, K14 ) k1/k4, ..., K1n ) k1/kn, we can solve for K1n as follows:
K1n )
Cstd1Astdn CstdnAstd1
(8)
(B) Calculation of Sample Composition. The total percent copolymer T (in mole percent or weight percent) composition of the polymer sample is given by
T ) Csam1 + Csam2 + Csam3 + Csam4 + ..., + Csamn
(9)
where Csam1, Csam2, Csam3, ..., Csamn are concentrations (in mole percent or weight percent) of components 1, 2, 3, ..., and n. The T is not required to equal 100%, which is often the case when the polymer sample contains a component that is not part of the copolymer. The corresponding experimentally determined peak area is expressed as Asamn for component n Rearrangement of eq 8, employing the polymer sample variables:
Csamn )
Csam1Asamn K1nAsam1
(10)
Substitution of eq 10 into eq 9 yields n
T ) Csam1 +
Csam1Asami
∑K i)2
1iAsam1
(11)
(7)
where kn is defined as the Py-GC response factor for each component. The proportionality constants, which are used to relate each monomer component in the polymer sample to a specific pyrolysis fragment, are generated by the following procedure. Equation 7 is expanded such that
Cstd1 ) k1Astd1
Then solving for Csam1,
T
Csam1 )
n
1+
Asami
∑K i)2
(12)
1iAsam1
Cstd2 ) k2Astd2 Cstd3 ) k3Astd3 l Cstdn ) knAstdn Then the ratio of these expressions are taken relative to C1 3034
Analytical Chemistry, Vol. 68, No. 17, September 1, 1996
With the known value of the total percent copolymer (in mole percent or weight percent) T, the K’s determined from the standard, and Csam1 calculated from eq 11, the rest of the components of the polymer sample can be determined by using eq 10. Knowing the standard composition and measuring the
standard and unknown sample monomer peak intensities, these equations can be used to calculated the composition for the unknown. EXPERIMENTAL SECTION Sample Preparation. (1) Styrene/Methyl Methacrylate Copolymer Resins. The homogeneous STY/MMA copolymer resins (labeled as samples A-G) were produced in the laboratory by radical polymerization using bulk processes as described in a standard textbook.13 (2) Styrene/Methyl Methacrylate Emulsion Polymer. The homogeneous STY/MMA emulsion polymers (sample I) and blended STY/MMA emulsion polymer (sample J) were synthesized and synthesized/blended in the laboratory based on procedures given in an experimental polymer textbook.14 The structured polymer (sample H) was synthesized in a similar way except monomers were fed in series. Py-GC and Py-GC/MS. Samples of polymer pellets (∼300 µg) were deposited in a quartz tube and then put into a Pt coil. Samples of emulsion polymer (∼3 µL) were deposited on the Pt ribbon and equilibrated for 10 min in a 250 °C interface connected to the injection port of a HP5890 gas chromatograph equipped with a flame ionization detector (FID). Samples were pyrolyzed (CDS 120 Pyroprobe Pt coil) at a set temperature of 700 °C (actual temperature is 715°C based on the factory calibration). The coil was heated to 700 °C at 20 °C/ms and held at 700 °C for a 20 s interval. The pyrolysis products were split in the 250 °C injection port, with 10 psi head pressure, 30:1 split ratio, separated on a fused-silica capillary column (J&W DB-5, 30 m × 0.25 mm i.d., 1.0 µm film) using a linear temperature program (40 °C/4min, 20 °C/min, to 120 °C/10 min, and then 20 °C/min ramp to 300 °C/23 min) and detected by a FID. In the GC/MS system, the FID was replaced with a VG Trio-1 mass spectrometer. The output from GC was transferred through a transfer line (280 °C) to the ion source of the mass spectrometer. Electron ionization mass spectrum was obtained every second over the mass range of 29-500 Da. Methane was used as reagent gas in the chemical ionization mode. Test of Temperature Dependence and Reproducibility. A series of pyrolysis runs of a STY/MMA copolymer in the temperature range from 500 to 900 °C were performed. The 700 °C pyrolysis temperature was chosen based on optimized yield of monomer and trimers for both methyl methacrylate and styrene. The reproducibility of the pyrolysis data was always a concern when the technique was applied to any kind of quantitative study. Based on eight consecutive runs of a STY/MMA copolymer, the normalized monomer and trimer peak intensities show a relative standard deviation below 1%, which demonstrates the reliability of the pyrolysis method applied to the analysis of STY/MMA copolymers. RESULTS AND DISCUSSION Figure 1 shows the typical pyrogram of a 39:61 (wt %) STY/ MMA homogeneous copolymer. The identification of all dimer and trimer peaks was accomplished by comparing chromatogram (13) Braun, D.; Cherdron, H.; Kern, W. Praktikum der makromolekularen organischen Chemie, 3rd. ed.; Huthig Verlag: Heidelberg, 1979; p 230. (14) Collins, E. A.; Bares, J.; Billmeyer, F. W., Jr. Experiments in Polymer Sciences; John Wiley & Sons: New York, 1973; pp 337-45.
Figure 1. Typical pyrogram of a STY/MMA (39:61 by wt %) copolymer.
retention times with a literature chromatogram15 as well as by comparing mass spectra obtained from Py-GC/MS in the EI mode and CI mode. The distinction of hybrid trimer peaks of SSM (styrene-styrene-methyl methacrylate) and SMS (styrenemethyl methacrylate-styrene) was accomplished by comparing chromatograms of STY/MMA homogeneous copolymer and STY/ MMA alternating copolymer.15 Table 1 lists the reference composition, calculated composition and number-average sequence length for all different compositions of STY/MMA homogeneous copolymers (labeled as samples A-G). Figure 2 shows those pyrograms of different compositions of STY/MMA homogeneous copolymers. The relative intensities of pure dimer, trimer and hybrid dimer, trimer peaks supplied a very good indication of relative monomer abundance. All the calculated compositions were in a good agreement with the NMR results. The number-average sequence length of styrene is calculated by using eq 1, which involves the styrene dimer and the hybrid STY/MMA dimer (Tables 1 and 2). The number-average sequence length of methyl methacrylate is calculated with eq 6, which utilized the number-average sequence length of styrene and the composition of the STY/MMA copolymer. In order to relate the dimer and trimer peak areas to the diad and triad distribution in the polymer, the response factor of Py-GC has to be determined. In this study, a STY/MMA copolymer (polymer E) was studied both by Py-GC and by 13C NMR spectrometry. The composition, as well as number-average sequence length for styrene and methyl methacrylate, was obtained by 13C NMR as a reference standard to calculate the corresponding response factor in Py-GC experiments. The response factor determined for styrene dimer was 1.0, and the response factor for hybrid STY/ MMA dimer was 3.0. The other structure-related terms such as grouped numberaverage sequence length and the percentage of grouped monomer4,5 were also calculated. In order to calculate the grouped number-average sequence length and the percentage of the grouped monomer, the triad distribution, MMM and MSM, had to be calculated based on the other triad distribution values and the number-average sequence length. Again, similar to number(15) Tsuge, S.; Ohtani, H. Pyrolysis Gas Chromatography of High Polymers Fundamentals and Data Compilation; Techno-System: Tokyo, 1989; pp 104-7.
Analytical Chemistry, Vol. 68, No. 17, September 1, 1996
3035
Table 1. Number-Average Sequence Length for Different Compositions of STY/MMA Narrow Composition Distribution Standard Polymers from Pyrolysis Gas Chromatography and 13C NMR Methodsa sample ID
std wt % S std wt % M wt % S wt % M N(S) N(M) %G(S) %G(M) GN(S) GN(M)
A
B
C
D
E
80b
68b
62b
50b
39b
20a 79.2 20.8 5.95 1.62 93 40 9.7 22.0
32a 69.0 31.0 3.86 1.80 89 51 5.9 8.2
38a 62.9 37.1 3.23 1.98 87 61 4.9 6.3
50a 50.2 49.8 2.31 2.38 81 70 3.4 5.9
61b 40.6 59.4 2.08 3.17 81 83 2.8 5.8
E 39 61 39 61 2.06 3.00 81 83 2.7 5.2
F 30 b 70 b 30.2 69.8 1.64 3.93 65 90 2.5 5.9
G 19 c 81c 18.5 81.5 1.34 6.17 47 96 2.2 7.9
a M, methyl methacrylate; S, styrene; N(S), N(M), number-average sequence length; %G(S), %G(M), percentage of grouped monomer units; GN(S), GN(M), number-average sequence length of grouped monomer units. n Composition data from 13C NMR results. c The composition data from synthesis receipt.
Table 2. Composition and Number-Average Sequence Length Data for Structured (H), Randomly Distributed (I), and Blend (J) Polymers from Pyrolysis Gas Chromatography Methoda sample ID
std wt % S std wt % M wt % S wt % M N(S) N(M) %G(S) %G(M) GN(S) GN(M) a See footnote a 13C NMR results.
Figure 2. Pyrograms (region of interest) of several STY/MMA copolymers with different composition.
average sequence length calculation, the 13C NMR results were used to calculate the response factors for different trimers. All trimer response factors were set to 1.0 except for the hybrid trimer NSMM+MMS, which was assigned a value 4.0. This implies a low efficiency of pyrolysis degradation or instability under the Py-GC conditions, The response factor, which is the combination of pyrolysis efficiency and FID response, is very important in the use of dimer and trimer peak areas to determine the diad and triad distributions. The dimer or trimer peak area is not necessary directly equal to the corresponding diad or triad distribution. A good example, found in this study, is the peak area of SSM and MSS triads. In theory, the triad distribution of SSM and MSS should be equal 3036 Analytical Chemistry, Vol. 68, No. 17, September 1, 1996
H
I
82b
82b
18b 82.8 17.2 61.25 13.21 99 97 889.5 28.4
J
18b 82.8 17.2 6.75 1.45 92 39 13.0 5.5
in Figure 1 for definitions.
b
82b 18b 82.8 17.2 1580 342 100 100 4190 493
Composition data from
for the STY/MMA copolymer system, but the peak areas of SSM and MSS are not the same. This is due to the response factor difference causing the same triad distribution to have slightly different peak areas. In the calculation of the grouped number-average sequence length and the percentage of the grouped monomer of styrene, the triad distribution term SSM and MSS must be used. Because the trimer peak areas of SSM and MSS are not equal, it is necessary to determined a consistent way to treat the sum of the triad term, NSSM+MSS. In this study, instead of applying a separate response factor to make SSM peak area equal to MSS, the peak areas of SSM and MSS have been added together without any correction to represent the triad distribution, NSSM+MSS. Three STY/MMA emulsion copolymers, which have the same compositions but different structures (structured polymer H, homogeneous polymer I, and blended polymer J), were the best examples to demonstrate the importance of structure exploration using Py-GC. The composition of these three polymers are 82.8 st % styrene and 17.2 wt % methyl methacrylate. Figure 3 shows the relative peak intensities of the dimers and trimers for these three copolymers. It is clearly demonstrated that all hybrid dimer and trimer peaks in polymer H are much less intense than polymer I but still can be seen. This is a simple indication that the distribution of different kinds of monomers in the polymer chain is limited. The polymer chain must exist as separate domains of
polymer degradation when a methyl group attaches to one of the vinyl carbons. The monomer is always the most favorable degradation product. In this situation, there is no detectable dimer or trimer available for any structure interpretation for Py-GC. The ultimate way to obtain structure information is to rely on the partially available dimer and trimer peaks and the composition, as demonstrated in this study. CONCLUSIONS
Figure 3. Pyrograms (region of interest) of structured (H), homogeneous (I), and blended (J) polymers.
different monomers with minimum interface area. For blended polymer J, all the hybrid dimer and trimer peaks do not exist because there is no repeat hetero units in the blend. The difference between polymer H (structured) and polymer J (blended) is related in these very low intensity hybrid dimer and trimer peaks. Table 2 shows the number-average sequence length and other structure-related terms for these three polymers. The structural differences between these three emulsion polymers can be seen by the large difference in the number-average sequence for both styrene and methyl methacrylate. The grouped number-average sequence length and the percentage of the grouped monomer strongly indicated that the polymer H has a separate domain of polymerized styrene and methyl methacrylate, but polymer J has entirely separate polymeric monomers. For copolymers containing monomers such as R-methylstyrene and aliphatic methacrylate, there is a unique phenomenon of
It was not possible to determine the number-average sequence length of STY/MMA copolymer systems using only the dimer and trimer peak distribution by pyrolysis gas chromatography because MMA dimer and MMA trimer fragments are not generated in detectable quantities. By determination of the numberaverage sequence length of styrene (which was calculated from dimers) and the composition (which is calculated from monomers), the number-average sequence length of methyl methacrylate was calculated. The structure information was further quantified by deriving the percent of grouped monomers and the number average length of grouped monomers. This method could be extended to other copolymer systems in which one of the monomers does not produce stable dimer and trimer fragments but gives stable monomer peaks. This class of polymers includes R-methylstyrene and aliphatic methacrylates. This development extends the structural study of copolymers by pyrolysis gas chromatography not only to the well-behaved copolymer system but also to those copolymer systems that produce incomplete dimer and trimer fragments.
Received for review February 28, 1996. Accepted June 10, 1996.X AC960196E X
Abstract published in Advance ACS Abstracts, August 1, 1996.
Analytical Chemistry, Vol. 68, No. 17, September 1, 1996
3037
