научная статья по теме ANALYSIS OF RAPID METHOD FOR THE QUALITY INDEX OF PHENOLIC RESINS AND BENEFIT Общие и комплексные проблемы технических и прикладных наук и отраслей народного хозяйства

Текст научной статьи на тему «ANALYSIS OF RAPID METHOD FOR THE QUALITY INDEX OF PHENOLIC RESINS AND BENEFIT»

Тепловые методы

УДК 620.179.13

ANALYSIS OF RAPID METHOD FOR THE QUALITY INDEX OF PHENOLIC RESINS AND BENEFIT

Y. Dong*1-2, S.Y. Qu1 1 School of management Harbin institute of technology, Harbin, China, 150001 2 Financial Office of Heilongjiang Province, Harbin, China, P.O. Box: 410, Harbin 150001, People's Republic of China

Abstract. The traditional analytical methods used to determine phenolic resin are slow and produce environmentally unfriendly waste. In this work, Near-infrared (NIR) spectroscopy has been applied for resin content of phenolic resins. The partial least square regression was used to develop the calibration model of the resin content. 8 samples were withdrawn at different time stages for analysis with the proposed quantitative models; the data thus obtained were compared with those provided by reference methods. The maximal predicted error and the standard deviation of the differences were 0.96 and 0.51 % for resin content. The results of the paired t-test revealed that there was no significant difference between the NIR method and the reference method. NIR spectroscopy is an effective choice for the accurate, expeditious analyzing quality of phenolic resin.

Key words: NIR spectroscopy; Partial least square regression; phenolic resin.

АНАЛИЗ ВОЗМОЖНОСТЕЙ БЫСТРОГО МЕТОДА КАЧЕСТВЕННОЙ ОЦЕНКИ ФЕНОЛЬНЫХ СМОЛ

Й. Донг1'2, С.Й. Ку1 1 Школа управления Харбинского института технологий, Харбин,

Китай, 150001

2 Финансовый офис области Хэйлунцзяна, Харбин, Китай 410, Харбин 150001, Китайская народная республика

Традиционный аналитический метод, используемый для определения содержания фенольных смол, медленный и производит экологически опасные отходы. В работе инфракрасная спектроскопия была применена для определения состава фенольных смол. Для разработки модели калибровки состава использовали метод наименьших квадратов. На различных временных стадиях были взяты 8 образцов для анализа с целью разработки количественных моделей; данные, полученные таким образом, сравнивали с данными, полученными известными методами. Максимальная ошибка и стандартное отклонение составили 0,96 и 0,51 % по составу. Результаты парных испытаний показали, что не было никакой значительной разницы между методом инфракрасной спектроскопии и известным методом. Инфракрасная спектроскопия — эффективный выбор для точного, быстрого анализа качества фенольных смол.

Ключевые слова: спектроскопия, линейная и квадратичная регрессия, фенольная смола.

1. INTRODUCTION

Phenolic resin is widely used as an industrial material because of its good heat resistance, electrical insulation, dimensional stability, and chemical resistance [1—2]. The resin content is the important index of the phenolic resin; it will influence the properties of composite products. The analytical methods typically used for this purpose are based on the classical techniques recommended. Such methods are usually time-consuming, use large amounts of reagents and produce waste. In fact, analysis times can range from a few minutes to several hours; this lengthens production cycles, raises costs and lowers outputs.

In recent years, the instrumental analytical techniques most frequently used to identify and characterize resins [3—5], which include mass spectrometry (MS), nuclear magnetic resonance (NMR), iR spectroscopy, gas chromatography (GC) and high performance liquid chromatography (HPLC). However, these methods cannot measure quality of the resin fleetly, non-destructively and contiguously.

88

H. aoht, C.H. Ky

Near infrared (NIR) spectroscopy is one of the most effective alternatives in this context. It is a rapid and nondestructive method for the simultaneous measurement of different constituents in various products. NIR spectrum contains information related polymer properties such as composition, conformation and crystallinity, therefore it can be widely applied for various polymer analyses in conjunction with chemometric calibration method [6—10]. The aim of the study is to develop an analysis method for the instantaneous prediction the resin content of the phenolic resin by diffuse reflection near-infrared spectroscopy. Calibration model of the resin content was developed by partial least squares (PLS), which were used to analyze unknown samples.

2. EXPERIMENTAL 2.1. Non-contact analyzing equipment

The quality analyzing equipment of the phenolic resin was shown in Fig. 1, which consisted of an FT-NIR spectrometer (Bruker Co., Germany), a bracket

Fig. 1. Sketch of analyzing the resin content of phenolic resin:

1 — NIR spectrometer; 2 — bracket of spectrometer; 3 — computer; 4 — bracket; 5 — gided metal plate; 6 — phenolic resin solution; 7 — beam.

with a gilded metal plate and a computer. Non-contact spectrum instrument was assembled resin up 17 cm. A gilded metal plate was placed under the resin in order to enhance the diffuse reflectance effect. The computer could be used to control the spectrometer.

2.2. Spectroscopy techniques

NIR spectral data was obtained by a non-contact scanning method for the phenolic resin, the light from the sources was focused on the resin. Then the diffuse reflectance spectra from the resin were recorded by the spectrometer. The NIR spectrometer was operated in the near infrared region from 12.000 to 4000 cm1. The resolution of the spectra is 8 cm1 and the average scanning times is 8, which were best according to spectrum quality and model result. The instrument was controlled via its bundled software (Bruker Co., Germany).

2.3. Reference methods of the resin content

The test specimen parts W weighed separately to the nearest 0.0001 g to obtain the initial weight W1. Part W1 was placed in an oven at 160 °C for 15 min,

cooled in a desiccator for 30 min, and immediately weighed to obtain the weight W2. The resin content (R %) was calculated as follows:

R % = (W1 - W2)/W1 x 100 (1)

3. RESULTS AND DISCUSSION 3.1. NIR spectra information of materials

The resin system mostly contains phenolic resin, phenol, ethanol and byproducts. The spectra of phenolic resin solution were shown in Fig. 2, which may be interpreted by assignment of the bands to overtones and combination of fundamental vibrations involving hydrogenic stretching and deformation modes. In

Fig. 2. NIR spectra of modified phenolic resin solution.

Fig. 2, the combination band of CH stretching and deformation mode given by the benzene groups was noticed in the region 4386—4255 cm-1. The band situated at approximately 5904 cm-1 was related to the first overtone of the symmet-

Fig. 3. First derivative and second derivative near infrared spectra of phenolic resin solution.

ric and asymmetric CH2 stretching vibrations, the 5936 cm-1 vibrational peak was associated with the aromatic CH stretching vibrations. The second overtone bands of the fundamental aliphatic CH2 stretching vibrations were noticed at

90

H. aoht, C.H. Ky

8800—8400 cm-1. The combination band of C=O stretching mode at 4434 cm-1. The alcohol and phenol groups gave a first overtone band of OH stretching mode and OH deformation mode at 6971 cm-1, combination band of OH stretching mode at 5190 cm-1. In order to eliminate variations in offset or different linear baselines and instrument noise, to ensure a good correlation between the spectral data and the concentration values, first derivative and second derivative spectral pretreatments were used, the result was showed in Fig. 3. From the above analysis, NIR spectra contained abundant information of samples, and a multivariate calibration method based on PLS could be used to obtain relevant analytical information and determine the quality index.

3.2. Quantitative analysis models

49 samples was developed the calibration model of the resin contents, the maximal resin content was 68.12 %, the minimum resin content was 64.66 %, and the mean value was 67.95 %. The partial least square regression was used to develop the calibration model of the resin content. In process of the developing model, first derivative pre-processing routines was used, the spectral regions were in 11990—7490 and 6100—4200, 6 PLS factors were given; full cross-validation was applied to optimize the calibration models. In the evaluation indexes of the calibration model, the determination coefficients (R2) was 0.95, root mean square error of calibration (RMSEC) was 0.19.

Table 1

The reliability evaluation of NIR-predicted versus reference values that was used to calculate the student's i-test (a = 0.05) for paired values

Resin content (%)

No

Actual Predicted Error

1 67.33 67.12 0.21

2 67.58 67.02 0.52

3 67.02 66.86 0.16

4 66.92 67.88 -0.96

5 67.41 66.94 0.56

6 67.34 67.76 -0.42

7 66.28 66.49 -0.21

8 66.71 66.95 -0.24

S D % 0.51

t experiment 0.13

t critical 2.36

In order to assess the predictive ability of developing model, 8 samples were collected randomly and analyzed by the reference method. The NIRS method and reference method were compared for resin content Student's t-test (a = 0.05) for paired values. In tab. 1, it is observed that t experiment value of the resin content was less than t critical values. The levels of significance obtained were 0.05, the result showed that there was no significant difference between the NIR method and the reference method, thus the NIR technique was reliable for the measurement of the resin quality.

3.3. Benefit analysis

The results of NIRS method and reference method were compared in tab. 2, it could be observed that the NIRS was a very rapid, noncontact and non-destruc-

tive method for the resin content of phenolic resin. Moreover, NIR method was the lowers analytical costs; it can economize the materials, reduce the production cycles and increase the outputs of the phenolic resin.

Table 2

Difference of NIR method and reference method

Reference method NIR method

Analysis of samples Destructive and contact Nondestructive and noncontact

Equipment Oven, balance, Beaker No

Process of analysis a

Для дальнейшего прочтения статьи необходимо приобрести полный текст. Статьи высылаются в формате PDF на указанную при оплате почту. Время доставки составляет менее 10 минут. Стоимость одной статьи — 150 рублей.

Показать целиком