Statistična analiza rezultatov DSC meritev granuliranega recikliranega PET z metodo glavnih osi

Klenovšek, Žan (author), Čelan Korošin, Nataša (mentor)

V diplomskem delu sem raziskoval termične lastnosti granuliranega recikliranega polietilentereftalata (PET) različnih izvorov z uporabo diferenčne dinamične kalorimetrije (DSC) z namenom ovrednotiti medsebojne razlike med vzorci, preučiti vpliv termične zgodovine materiala ter preveriti uporabnost kemometrijskih metod za razvrščanje polimerov in prepoznavanje odstopajočih vzorcev. Meritve sem izvajal na diferenčnem dinamičnem kalorimetru Mettler Toledo DSC 5+ po cikličnem programu segrevanja nad temperaturo tališča in ohlajanja pod temperaturo steklastega prehoda. Pridobljene DSC krivulje sem obdelal s programsko opremo STARe Evaluation Software ter karakteriziral termične pojave: steklasti prehod, kristalizacijo in taljenje. Pridobljene fizikalne parametre kristalizacijsko entalpijo (∆Hc), talilno entalpijo (∆Hm), spremembo toplotne kapacitete (∆Cp,hg) ter karakteristične temperature (Tonset,c, Tonset,m, Tonset,hg, Tmid,ISO,hg in Tpeak,m, Tpeak,c) sem statistično vrednotil z metodama Pearsonove korelacijske analize in analize glavnih osi (PCA) v programu OriginPro v treh zaporednih krogih, pri katerih sem izločal odstopajoče vzorce, osamelce. Rezultati so pokazali, da imajo reciklirani vzorci kljub različni zgodovini predelave podobne osnovne termične lastnosti kot referenčna PET folija. Večja odstopanja so bila opazna pri pojavu hladne kristalizacije in vrednostih kristalizacijske entalpije. Z analizo PCA sem v tretjem krogu s prvima dvema glavnima osema (PC1 in PC2) pojasnil 61,69 % skupne variance podatkov. Pearsonova korelacijska matrika je razkrila zelo močno pozitivno korelacijo (povezanost) (angl. Correlation) (r = 0,96) med začetno temperaturo steklastega prehoda pri 2. gretju (Tonset,hg) in temperaturo srednje točke steklastega prehoda (Tmid,ISO,hg) ter srednje negativno korelacijo (r = −0,55) med spremembo toplotne kapacitete (∆Cp,hg) in kristalizacijsko entalpijo med hlajenjem (∆Hc). Diagrami razporeditve so potrdili jasno združevanje v gruče, zlasti pri barvno definiranih vzorcih podjetja rPET InWaste ter vzorcih s podobno stopnjo kristaliničnosti. Statistična analiza je omogočila prepoznavanje neustreznih vzorcev, določitev medsebojnih povezav med termičnimi parametri in razvrščanje vzorcev v skupine glede na njihove podobne lastnosti ter izvor predelave.

• Language: Slovenian
• Keywords: diferenčna dinamična kalorimetrija (DSC), granuliran recikliran polietilen tereftalat (PET), metoda glavnih osi (PCA), kemometrija
• Work type: Bachelor thesis/paper
• Organization: FKKT - Faculty of Chemistry and Chemical Technology
• Year: 2026
• PID: 20.500.12556/RUL-186624
• Publication date in RUL: 03.09.2026

Secondary language

• Language: English
• Title: Statistical Analysis of DSC Measurement Results for Granulated Recycled PET Using Principal Component Analysis

Abstract

In this diploma thesis, the thermal properties of granulated recycled polyethylene terephthalate (rPET) from various sources were investigated using differential scanning calorimetry (DSC) to evaluate differences between samples, study the effect of thermal history, and test the applicability of chemometric methods for polymer classification and outlier identification. Measurements were performed on a Mettler Toledo DSC 5+ differential scanning calorimeter using a cyclic heating-and-cooling programme above the melting point and below the glass transition temperature. The DSC curves obtained were processed using STARe Evaluation Software to characterise thermal phenomena: glass transition, crystallisation, and melting. The extracted physical parameters, crystallisation enthalpy (∆Hc), melting enthalpy (∆Hm), heat capacity change (∆Cp,hg) and characteristic temperatures (Tonset,c, Tonset,m, Tonset,hg, Tmid,ISO,hg in Tpeak,m, Tpeak,c) were statistically evaluated using Pearson correlation analysis and principal component analysis (PCA) in OriginPro across three consecutive rounds of outlier elimination. The results showed that, despite their different processing histories, the recycled samples exhibited basic thermal properties similar to those of the reference PET film. More significant deviations were observed in the occurrence of cold crystallisation and in the crystallisation enthalpy values. In the third round of PCA, the first two principal components (PC1 and PC2) accounted for 61.69% of the total variance in the data. The Pearson correlation matrix revealed very strong positive correlation (r = 0.96) between the onset glass transition temperature during the second heating (Tonset,hg) and the midpoint glass transition temperature (Tmid,ISO,hg), as well as a moderate negative correlation (r = −0.55) between the heat capacity change (∆Cp,hg) and the crystallisation enthalpy during cooling (∆Hc). Score plots confirmed distinct clustering, particularly among colour-defined samples from rPET InWaste and samples with a similar degree of crystallinity. Statistical analysis enabled the identification of outlying samples, the determination of relationships among thermal parameters, and the classification of samples into groups based on their shared properties and processing origin.

• Keywords: differential scanning calorimetry (DSC), granulated recycled polyethylene terephthalate (PET), principal component analysis (PCA), chemometrics