J Polym Environ (2006) DOI 10.1007/s10924-006-0001-z
4
Optical Properties of Polylactides Matthew H. Hutchinson Æ John R. Dorgan Æ Daniel M. Knauss Æ Sukhendu B. Hait
F
3
ORIGINAL PAPER
PR OO
1 2
5
32 33 34 35 36 37
Keywords Polylactides Æ PLA Æ Poly(lactic acid) Æ Index of refraction Æ Cauchy parameters Æ Dispersion
38 39
Introduction
40
Polylactides are environmentally benign and degradable polymers made from renewable resources [1]. They have physical properties that make them useful for fibers, packaging, and other applications traditionally dominated by petroleum-based resins. The recent advent of large-scale commercial production of these materials makes characterization of their fundamental properties highly desirable. Although the general literature on polylactides is extensive, recent reviews [2, 3] point out that the optical properties (i.e. the index of refraction and its dependence on wavelength) have not been reported. The structure of PLA is unique in that the polymer can contain enantiomers of both L- and D-lactic acid as repeat units. Generally, PLAs are predominantly L in stereochemical composition, however, varying levels of the D-stereoisomer can be present. Once the optical purity falls below about 80%, PLA becomes fully amorphous rather than semicrystalline. To date, no study has captured a systematic description of PLA optical properties across a broad range of stereooptical composition (i.e. across a wide range of L-content, the commercial materials being usually high (>90%) in L-stereochemical centers).
41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 61 62
D
the Cauchy coefficients increases with increasing wavelength, and ranges from 1.1% at 300 nm to 0.1% at 1300 nm. The final useful result of the study is that the index of refraction of PLA within the range of wavelengths from 300 nm to 1300 nm may be described by n(k) = (1.445 – 0.00075) + (4892 – 143) nm2/k 2.
RR
EC
TE
Abstract Polylactides (PLAs) have been known for several decades but have only recently gained commercial significance as a leading environmentally benign plastic available from renewable resources. Accordingly, it is highly desirable to understand the optical properties of these materials. Optical properties are important in dyeing operations for textile applications and in various packaging applications where coloring or clarity is desirable, however, a description of optical properties is not available in the literature. In this study, ellipsometric measurements were made on a well-characterized set of homopolymers and copolymers spanning a wide range stereoisomer proportions (L-content). Data were collected and analyzed within the framework of the well-known Cauchy model for the index of refraction. No statistical difference is found for the Cauchy parameters for the various optical compositions. The data for all samples can be accurately described using values of Cauchy coefficients of A=1.445 – 7.529·10)4 and B=4.8916·103 – 1.426 · 102 nm for ‘B’. It was found that over the light wavelengths from 300 to 1300 nm, the index of refraction for PLA decreased from 1.499 to 1.448. The uncertainty in the index of refraction calculated from
CO
9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31
Received: j / Accepted: j
Springer Science+Business Media, Inc. 2006
M. H. Hutchinson Æ J. R. Dorgan (&) Department of Chemical Engineering, Colorado School of Mines, Golden, CO 80401, USA E-mail:
[email protected]
UN
6 7 8
D. M. Knauss Æ S. B. Hait Department of Chemistry and Geochemistry, Colorado School of Mines, Golden, CO 80401, USA M. H. Hutchinson Department of Chemical Engineering, Cambridge University, Cambridge, UK
Journal : 10924 PIPS No.
: xxxxxxx
MS Code : 1
Dispatch : 17-2-2006
Pages : 6
h LE 4 CP h
h TYPESET 4 DISK h
2
J Polym Environ (2006)
Materials and methods
CO
RR
EC
TE
The materials investigated were synthesized in the laboratories of the Colorado School of Mines. All polymerizations were performed using the following procedure, with different ratios of L- to D-lactide and amounts of benzyl alcohol. Briefly, L-lactide and D-lactide (Purac Biochem., Holland) were crystallized from toluene. Tin-2ethylhexanoate (Aldrich Chemical Co., USA) was distilled under vacuum before use. Benzyl alcohol (Aldrich Chemical Co., USA) was dried with CaO and distilled under reduced pressure. Lactide (50 g, 0.347 mol) and Sn(Oct)2 (5.63 mg, 0.138 mmol) were added to a 50 mL round-bottom flask, dried under vacuum for 4 h, sealed under a rubber septum, and then purged with argon for 20 min. The required amount of benzyl alcohol to achieve a target molecular weight was added using a syringe. The polymerization reaction was performed by heating the vessel at 130 C for 12 h under an argon atmosphere. The relatively low polymerization temperature is used to avoid transesterification and chain scission reactions that increase polydispersity and can produce changes in stereochemical content. The resulting polymer was dissolved in dichloromethane, precipitated with a 10-fold excess of hexane, and dried under reduced pressure (1 torr) for 48 h. Weightaverage molecular weights obtained spanned the range from below 104 to over 106, and the range of enantiomer proportions (L:D) was from 100:0 to 50:50. Identification and specifications of the polymers studied are given in Table 1. Routine thermal characterizations of the various polymers of Table 1 were done using a Perkin-Elmer DSC-7 after calibration with indium. Results are collected in Table 1. Reported glass-transition temperatures (Tg’s) are half-step-height values and melting points are peaks of Table 1 Experimental polylactidesa Sample SH-PLA-72b SH-PLA-14 SH-PLA-43 a
%L /%
D
UN
70 71 72 73 74 75 76 77 78 79 80 81 82 83 84 85 86 87 88 89 90 91 92 93 94 95 96 97 98 99 100 101 102
F
69
melting endotherms in second-heating scans at 10 C/min after cooling from the melt to room temperature at the same rate. For the homopolymers (SH-PLA-72 is 100:0 in stereochemical composition), the tabulated Tg value refers to a specimen quenched as quickly as possible in the instrument from the melts to room temperature to suppress crystallization. Ellipsometric measurements were conducted utilizing a WVASE spectroscopic ellipsometer (J.A. Woolham, Lincoln, NB). Ellipsometry can be used to obtain thicknesses, interfacial roughness, domain sizes, and indices of refraction in thin films; it measures the change in the state of polarized light upon reflection [4]. Samples were studied using wavelengths ranging from 300 to 1300 nm and angle of incidence ranging from 65 to 75 degrees. The range of incident angles was chosen to maximize sensitivity in the measured ellipsometric parameters Y and D. Silicon wafers were utilized as substrates for the PLA films. The wafers were first carefully characterized using the ellipsometer to determine the thickness of the silicon dioxide layer. Subsequently, PLA was spin-coated from solution onto the wafers to form thin film samples. Two different solvents were used; tetrahydrofuran was used as the solvent for the PLA with L:D ratios of 50:50 and 80:20, and methylene chloride was used for PLA with L:D ratios of 100:0. For each of the PLA L:D ratios used, two different solution concentrations were made to produce two different film thicknesses. Solution concentrations were selected based on previous experience to produce a thickness of the PLA film of about 150 nm for the less concentrated solution and about 450 nm for the higher concentration solution. Spin-coating was conducted at 2000 rpm for 2 min. The thin film samples were subsequently annealed overnight in a vacuum oven (22 inches of water vacuum at 5600 feet of altitude) at 40 C to remove residual solvent. The Cauchy model was used to model the index of refraction n as a varying function of wavelength, k. The Cauchy model is given by Eq. (1),
PR OO
One of the predominant uses for PLA is in the creation of fibers that can then be used in fabrics. This use raises questions about the optical properties of the PLA during dyeing operations. The present study determines the index of refraction for PLA over a range of wavelengths including the visible spectrum.
D
63 64 65 66 67 68
100/0 80/20 50/50
10)3Mn
10)3Mw
Tg, C
Tm, C
77.7 70.40 84.00
120 103 172
59.5 51.5 52.0
177.4 – –
Molecular weights are from SEC in THF with light-scattering detection unless noted otherwise
b
Molecular weights from intrinsic viscosities in chloroform
Journal : 10924 PIPS No.
: xxxxxxx
MS Code : 1
nð kÞ ¼ A þ
B C þ k2 k4
ð1Þ
where n represents the index of refraction (a function of the wavelength) and the parameters A, B, and C are to be determined by experiment. The spectroscopic ellipsometer WVASE32 program was used to implement least squares regression to find the best fit of the experimental data by varying the film thickness, the cauchy parameters A and B, and the film non-uniformity. Based on the data fits, it was determined that it was appropriate to simply fix the Cauchy parameter C at zero. Films were also investigated using a using a Tencor P-10 Surface Profilometer. Deposited films were cut with a
Dispatch : 17-2-2006
Pages : 6
h LE 4 CP h
h TYPESET 4 DISK h
103 104 105 106 107 108 109 110 111 112 113 114 115 116 117 118 119 120 121 122 123 124 125 126 127 128 129 130 131 132 133 134 135 136 137 138 139 140
142 143 144 145 146 147 148 149 150 151 152 153
J Polym Environ (2006)
3
razor blade to create a groove and the polymer material to one side of the groove was carefully removed. A height versus distance profilometer trace was then collected to obtain film thickness. The thickness found using the profilometer was subsequently compared to that determined by ellipsometry as a means of independently verifying the film thicknesses measured by ellipsometry.
161
Results and discussion
162 163 164 165 166 167 168 169 170 171 172 173 174 175 176 177 178 179 180 181 182 183 184
To determine the Cauchy coefficients ‘A’ and ‘B’, the experimental data were modeled by varying the Cauchy coefficients, the film thickness, and the film uniformity. Our method for determining that the best-fit analysis is generating accurate results is to compare the results obtained for the thickness to measured values from profilometry. The percent difference between the thickness determined by ellipsometry and that found with the profilometer for the thinner films ranged from 0.1 to 7.6%, with an average percent difference of 3.9%. This is excellent agreement across independent measurements. The best-fit results coincide with the experimentally measured filmthickness, which validates the best-fit results. Results of the measurements on the L:D ratio of 50:50 sample are shown in Figs. 1 and 2. Figure 1 presents experimentally measured Y values over a range of wavelengths at a 70 angle of incidence. Also shown are the data fits and corresponding values. For this data set, the percent difference between the experimental Y values and those calculated from the best-fit model is on average 2.1% with a maximum percent difference of 8.1%. Figure 2 presents the D data—the percent difference between the experimental D values and those calculated from the best-fit
model is on average 3.4% with a maximum percent difference of 4.7%. Similar results are obtained for both the 80:20 and 100:0 samples. The index of refraction should be constant for each PLA structure and should not depend on the thickness of the film. However, performing ellipsometry measurements only does produce an apparent dependence on film thickness. In order to assess data reliability, two aspects were considered: the first being agreement with the film thickness as measured by profilometry and the second being the magnitude of the mean standard error (MSE) between the data and the fit to the Cauchy model. Lower MSE values and more accurate film thickness values were taken to imply better values for the Cauchy parameters. When the average thick film values for the Cauchy coefficients were used to describe the thin film data, the MSE increased from an average of 15.9 to 27.1 (an increase of 70%). The percent difference between the calculated thickness based on the ellipsometry and film thickness determined by profilometry increased from 4.16 to 5.04% (an increase of 21%). When the thin film Cauchy coefficients were used to describe the thick films, the MSE of the best fit increased by only 5% and the difference between the profilometry and ellipsometry determined film thicknesses was improved from 4.8 to 3.3%. That is, using the Cauchy coefficients derived from the thick film measurements for the analysis of the thinner films drastically increased both the MSE and the percent difference between the estimated and calculated film thickness. However, using the thin film Cauchy coefficients to analyze the thicker films only slightly increased the MSE and actually improved the agreement in film thickness. This analysis suggests that the data obtained from the thin film analysis is more accurate than the data obtained from the thicker films.
UN
CO
Fig. 1 Measurement and parametric fit of the ellipsometric Psi parameter as a function of wavelength for a PLA sample having a 50:50 L:D optical composition
RR
EC
TE
D
PR OO
F
154 155 156 157 158 159 160
Journal : 10924 PIPS No.
: xxxxxxx
MS Code : 1
Dispatch : 17-2-2006
Pages : 6
h LE 4 CP h
h TYPESET 4 DISK h
185 186 187 188 189 190 191 192 193 194 195 196 197 198 199 200 201 202 203 204 205 206 207 208 209 210 211 212 213 214 215 216 217 218
4
J Polym Environ (2006)
D
films. While every effort was made to exclude crystallization and while optical microscopy shows no birefringent regions within the films it is still possible that microscopic crystalline regions exist. If such microscopic regions are variable from sample to sample variation can be expected. Figure 4 shows the obtained values for the cauchy coefficient ‘B’ from the best-fit analysis for the three different L:D ratios. Again, the error bars for the L:D ratios of 50:50, 80:20, and 100:0 were calculated using respectively six, four, and four separate samples and represent one standard deviation in each direction. As with the ‘A’ parameters, there are no statistical differences between the ‘B’ Cauchy coefficients for samples with various L:D ratios. Once it was determined that there is no statistical significance of the ‘A’ or ‘B’ Cauchy coefficients with varying optical composition, regression was used to simultaneously
CO
Fig. 3 Average values for the A-parameter of the Cauchy model for varying optical compositions of PLA
RR
EC
TE
This is because the results are physically more accurate (as determined by measuring film thickness) and because the mean standard error in the best fit analysis is far better for the thinner films. Such a result is not unexpected as ellipsometry is best suited to thin films and the uniqueness of the data fitting becomes problematic as film thickness increases [4]. Figure 3 presents the results of the Cauchy model fitting results for the three different sets of samples with changing L:D ratios. The error bars represent one standard deviation in each direction, and are calculated for the L:D ratios of 50:50, 80:20, and 100:0 using six, four, and four different samples, respectively. Within the standard deviation of the test results, there is no statistical evidence that the cauchy coefficient ‘A’ value changes with changing L:D ratio of the PLA. The greater variation in this parameter for the 100:0 samples is attributed to possible crystallization within the
UN
219 220 221 222 223 224 225 226 227 228 229 230 231 232 233 234 235
PR OO
F
Fig. 2 Measurement and parametric fit of the ellipsometric Delta parameter as a function of wavelength for a PLA sample having a 50:50 L:D optical composition
Journal : 10924 PIPS No.
: xxxxxxx
MS Code : 1
Dispatch : 17-2-2006
Pages : 6
h LE 4 CP h
h TYPESET 4 DISK h
236 237 238 239 240 241 242 243 244 245 246 247 248 249 250 251 252
J Polym Environ (2006)
5
nðkÞ ¼ ð1:445 0:00075Þ þ ð4892 143Þ nm2 =k2
ð2Þ
Fig. 5 Index of refraction for PLA as a function of wavelength from a global determination of the Cauchy parameters across all optical compositions
Journal : 10924 PIPS No.
: xxxxxxx
MS Code : 1
D
275 276 277 278 279 280 281
Conclusions
282
Polylactides have physical properties that make them useful for fibers and packaging. In addition, they are environmentally benign and degradable polymers made from renewable resources. The recent advent of large-scale commercial production of these materials makes characterization of their fundamental properties highly desirable. Despite this, recent reviews point out that the optical properties (i.e. the index of refraction and its dependence on wavelength) have not been reported. This study provides this important information in a comprehensive fashion for this first time in the literature. Understanding
283 284 285 286 287 288 289 290 291 292 293
CO
RR
It is worth noting that at the wavelength of the sodium D line for which most tabulated solvent refractive indexes refer, evaluation of Eq. (2) gives nPLA(589.2)=1.459. Equation (2) represents the primary finding of this study.
UN
269 270 271 272 273 274
The change of the index of refraction with changing wavelength is called dispersion and this quantity is of interest in many technologies. The dispersion for PLA was calculated by taking the derivative of the Cauchy equation with respect to the wavelength. The results of this calculation are shown in Fig. 6. In this figure, the absolute value of the dispersion is shown.
TE
solve for the coefficients across all of the separate samples, regardless of the L:D ratios. This returned values of A=1.445 – 7.529·10)4 and B=4.8916·103 – 1.426·102 nm. Figure 5 presents the index of refraction of PLA as a varying function of wavelength calculated from Eq. (1) using these global values for the Cauchy coefficients. Over the range of wavelengths presently studied, the index of refraction decreases from 1.499 to 1.448 as the wavelength of the incident light increases from 300 to 1300 nm. The expected trend for the index of refraction is to decrease with increasing wavelength is observed. Inserting the determined Cauchy values into Eq. (1) gives the following useful expression for the index of refraction of PLA as a function of wavelength within the range of wavelengths from 300 nm to 1300 nm.
EC
253 254 255 256 257 258 259 260 261 262 263 264 265 266 267
PR OO
F
Fig. 4 Average values for the B-parameter of the Cauchy model for varying optical compositions of PLA
Fig. 6 Dispersion for PLA as a function of wavelength from a global determination of the Cauchy parameters across all optical compositions
Dispatch : 17-2-2006
Pages : 6
h LE 4 CP h
h TYPESET 4 DISK h
6
306 307 308
References
309
1. Hartmann MH (1998) In: Kaplan DH (ed) Biopolymers from renewable resources. Springer-Verlag, Berlin, pp 367–411 2. Auras RA, Harte B, Selke S (2004) Macromol Biosci 4:835–864 3. Garlotta D (2001) J Polym Environ 9:63–84 4. Azzam R, Bashara N (1987) Ellipsometry and polarized light. North-Holland, Amsterdam
310 311 312 313 314 315 316 317
PR OO
F
and B=4.8916·103 – 1.426·102 nm for ‘B’. This results in the useful expression for the index of refraction of PLA as a function of wavelength given by Eq. (2).
CO
RR
EC
TE
D
optical properties is important in dyeing operations for textile applications and in various packaging applications where clarity is desirable. Ellipsometric measurements were made on a comprehensive and well-characterized set of homopolymers and copolymers spanning a wide range stereoisomer proportions (l-content). Data were collected and analyzed within the framework of the well-known Cauchy model for the index of refraction. No statistical difference is found for the Cauchy parameters for the various optical compositions. The data for all samples can be accurately described using values of Cauchy coefficients of A=1.445 – 7.529·10)4
UN
294 295 296 297 298 299 300 301 302 303 304 305
J Polym Environ (2006)
Journal : 10924 PIPS No.
: xxxxxxx
MS Code : 1
Dispatch : 17-2-2006
Pages : 6
h LE 4 CP h
h TYPESET 4 DISK h
