Electronic Theses and Dissertation

Selama ini, kualitas minyak nilam dianalisis menggunakan gas chromatography-mass spectrometry (gc-ms) yang biayanya mahal dan membutuhkan waktu yang agak lama. sehingga diperlukan identifikasi yang cepat, mudah, praktis, dan murah. salah satunya metode yang sedang dikembangkan untuk menganalisis produk pertanian dan pangan yaitu teknologi near infrared reflectance spectroscopy (nirs). tujuan dari penelitin ini yaitu: mempelajari fitur spektrum nirs minyak nilam, minyak pencampur, dan minyak nilam campuran; membangun model pendugaan kadar patchouli alcohol (pa); membangun model pendugaan kadar bahan pencampur dalam minyak nilam; dan mengembangkan dan membandingkan beberapa pretreatment spektrum dalam mendukung ketangguhan model yang dibangun. bahan dan alat yang digunakan dalam penelitian ini antara lain: minyak nilam, minyak terpentin, minyak kelapa sawit, minyak kelapa, dan minyak keruing; micropipet, magnetic stirrer, rotary vacum evaporator, ft-nir thermo nicolet antaris ii tm, software the unscrambler 10.4, dan peralatan laboratorium lainnya. kandungan kimia dari hasil akuisisi spektrum nirs diterjemahkan dengan menggunakan analisis kemometrik. analisis kemometrik dalam penelitian ini dipergunakan untuk pengecekan data outlier, pretreatment spektrum, pembuatan model kuantitatif, dan validasi model. pengecekan data outlier menggunakan metode principal component analysis (pca) dan hotelling t² ellipse. pretreatment spektrum menggunakan metode mean normalization (mn), de-trending (dt), dan standard normal variate (snv). pembuatan model kalibrasi menggunakan metode principal component regression (pcr) dan partial least squares (pls); dan validasi model menggunakan metode independen validation. hasil penelitian menjelaskan bahwa teknologi nirs dengan metode kemometrika membuka peluang pengembangan standar mutu untuk minyak nilam. teknologi nirs mampu menduga molekul-molekul secara umum dalam minyak nilam. secara analisis spektrum, minyak nilam yang dicampurkan dengan minyak lain di atas 50% cenderung akan mengalami perubahan fitur spektrum. analisis lanjutan menggunakan kemometrik mendapatkan bahwa untuk pendugaan kadar pa minyak nilam, teknologi nirs berhasil mendapatkan model kalibrasi (rpd >2) namun masih memerlukan pengembangan lebih lanjut untuk model validasi. untuk pendugaan kadar bahan pemalsu dalam minyak nilam, teknologi nirs berhasil menduga kadar bahan pemalsu dalam minyak nilam terutama untuk model kalibrasi (rpd >3), namun masih memerlukan pengembangan lebih lanjut untuk model validasi. penambahan pretreatment spektrum pada model pendugaan kadar pa belum mampu meningkatkan ketangguhan model kalibrasi. sedangkan penambahan pretreatment spektrum pada model pendugaan kadar bahan pencampur cenderung meningkatkan ketangguhan model kalibrasi namun belum begitu memuaskan untuk peningkatan model validasi. hasil penelitian juga menjelaskan bahwa pencampuran minyak nilam dengan minyak keruing cenderung lebih sulit dideteksi dibandingkan dengan menggunakan minyak terpentin, minyak kelapa sawit, dan minyak kelapa. kata kunci: minyak nilam, nirs, pcr, pls.

So far, the quality of patchouli oil has been analyzed using Gas Chromatography-Mass Spectrometry (GC-MS), which is expensive and takes a long time. So fast, easy, practical and cheap identification is needed. One of the methods that is being developed to analyze agricultural and food products is Near Infrared Reflectance Spectroscopy (NIRS) technology. The objectives of this research are: to study the NIRS spectrum features of patchouli oil, blending oil, and mixed patchouli oil; building a model for estimating patchouli alcohol (PA) levels; building a model for estimating the levels of adulterants in patchouli oil; and developing and comparing several spectrum pretreatments to support the robustness of the models built. The materials and tools used in this research include: Patchouli oil, turpentine oil, palm oil, coconut oil, and gurjun oil; micropipette, magnetic stirrer, Rotary Vacum Evaporator, FT-NIR Thermo Nicolet Antaris II TM, The Unscrambler 10.4 software, and other laboratory equipment. The chemical content of the NIRS spectrum acquisition results was translated using chemometric analysis. Chemometric analysis in this research was used to check outlier data, spectrum pretreatment, create quantitative models, and model validation. Checking outlier data uses the Principal Component Analysis (PCA) and Hotelling T² Ellipse methods. Spectrum pretreatment uses Mean Normalization (MN), De-trending (DT), and Standard Normal Variate (SNV) methods. Creating a calibration model using the Principal Component Regression (PCR) and Partial Least Squares (PLS) methods; and model validation using independent validation methods. The research results explain that NIRS technology with the chemometric method opens up opportunities for developing quality standards for patchouli oil. NIRS technology is able to predict general molecules in patchouli oil. Based on spectrum analysis, patchouli oil mixed with other oils above 50% tends to experience changes in spectrum features. Further analysis using chemometrics found that for estimating the PA content of patchouli oil, NIRS technology succeeded in obtaining a calibration model (RPD > 2) but still requires further development for the validation model. For estimation levels of adulterants in patchouli oil, NIRS technology has succeeded in estimating levels of adulterants in patchouli oil, especially for the calibration model (RPD > 3), but still requires further development for the validation model. The addition of spectrum pretreatment to the model for estimating PA levels has not been able to improve robustness of the calibration model. Meanwhile, the addition of spectrum pretreatment to the mixing material content estimation model tends to increase the robustness of the calibration model but is not very satisfactory for improving the validation model. The research results also explain that mixing patchouli oil with gurjun oil tends to be more difficult to detect compared to using turpentine oil, palm oil, and coconut oil. Keywords: Patchouli oil, NIRS, PCR, PLS.