In this work, the important chromat

In this work, the important chromatographic factors
were selected and optimized by a central composite design experiment. The selection of factors for
optimization was based on preliminary experiments and
prior knowledge from literature, as well as certain
instrumental limitations. For instance, the mobile phase
pH was fixed at 3.0 as this could influence the
stability.13From preliminary experiments,the mobile
phase consistingof a methanol and triethylamine buffer
was employed in which the concentration of methanol
content was varied.14The mobile phase flow rate could
also moderately influence selectivity in the HPLC
analysis. Therefore, the key factors selected for the
optimization process were methanol concentration (A),
buffer molarity (B), and flow rate (C). Table 1 shows the
levels of each factors studied for finding out the optimum
values and responses. As can be seen in this table, the
ranges of each factor used were: methanol concentration
(60–70%), buffer molarity (15–25 mM), and flow rate
(0.7–0.9 ml/min).
As response variables, the capacity factor of
lamivudine (k1), the resolution between tenofovir and
efavirenz (Rs2, 3), and the retention time of efavirenz (tR3)
were chosen. All experiments were performed in
randomized order to minimize the effects of uncontrolled
variables that may introduce a bias on the measurements.
For an experimental design with three factors, the model
including linear, quadratic, and cross terms can be
expressed as:
Y = β0 + β1X1 + β2X2 + β3X3 + β12X1X2 + β13X1X3+β23X2X3
+ β11X21+ _β22X22+ β33X2
whereY is the response to be modeled, β is the
regression coefficient, and X1, X2 and X3 represent
factors A, B and C, respectively. To obtain a simple and
yet realistic model, the insignificant terms (P > 0.05) are
eliminated from the model through a ‘backward
elimination’ process. The statistical parameters obtained
from the ANOVA for the reduced models are given in
table 2.Since R2 always decreases when a regressor
variable is eliminated from a regression model, in
statistical modeling the adjusted R2, which takes the
number of regressor variables into account, is usually
selected. Adjusted R2was defined as: 1- SSE (n-p) / SST
(n-1) = 1-(n-1)/(n-p)x(1-R2)

In the present study, the adjusted R2 was well within the
acceptable limits of R2 .0.80 15 which revealed that the
experimental data shows a good fit with the second-order
polynomial equations. For all the reduced models, P value of

Hasil (Bahasa Indonesia) 1:

Dalam karya ini, faktor-faktor penting kromatografidipilih dan dioptimalkan oleh percobaan desain komposit pusat. Pemilihan faktoroptimasi didasarkan pada eksperimen Pendahuluan danpengetahuan sebelumnya dari sastra, serta tertentuinstrumental keterbatasan. Sebagai contoh, fase mobilepH ditetapkan pada 3.0 seperti ini dapat mempengaruhiStability.13From awal percobaan, mobiletahap consistingof metanol dan Trietilamina bufferdipekerjakan di mana konsentrasi metanolkonten adalah laju aliran varied.14The mobile tahap bisajuga cukup mempengaruhi selektivitas di HPLCanalisis. Oleh karena itu, faktor kunci yang dipilih untukproses optimasi yang metanol konsentrasi (A),Molaritas penyangga (B), dan laju aliran (C). Tabel 1 menunjukkantingkat masing-masing faktor belajar untuk mencari tahu yang optimalnilai-nilai dan tanggapan. Seperti dapat dilihat dalam tabel ini,Rentang setiap faktor yang digunakan adalah: metanol konsentrasi(60-70%), buffer Molaritas (15-25 mM), dan arus tingkat(0.7-0,9 ml/min).Sebagai respon variabel, faktor kapasitaslamivudine (k1), resolusi antara tenofovir danefavirenz (Rs2, 3), dan waktu penyimpanan efavirenz (tR3)dipilih. Seluruh eksperimen dilaksanakan diurutan acak untuk meminimalkan efek tidak terkendalivariabel yang dapat memperkenalkan bias pada pengukuran.Untuk rancangan percobaan dengan tiga faktor, modeltermasuk linear, kuadrat, dan salib persyaratan dapatdinyatakan sebagai:Y = Β0 + Β1X1 + Β2X2 + Β3X3 + Β12X1X2 + Β13X1X3 + Β23X2X3+ Β11X21 + _Β22X22 + Β33X2whereY adalah respon yang akan dibuat modelnya, βregresi dengan koefisien, dan X1, X2 dan X3 mewakilifaktor-faktor A, B dan C, masing-masing. Untuk memperoleh sederhana danmodel namun realistis, syarat-syarat tidak signifikan (P > 0,05) adalahdihilangkan dari model melalui ' mundurpenghapusan ' proses. Parameter statistik yang diperolehdari ANOVA untuk model berkurang diberikan dalam2. meja R2 karena selalu menurun ketika regressorvariabel telah dihapuskan dari model regresi, dalamStatistik pemodelan R2 disesuaikan, yang membawajumlah variabel regressor ke rekening, yang biasanyadipilih. Disesuaikan R2was didefinisikan sebagai: 1-SSE (n-p) / SST(n-1) = 1-(n-1)/(n-p)x(1-R2)Dalam penelitian ini, R2 disesuaikan adalah baik dalambatas-batas yang wajar R2 15.0.80 mengungkapkan bahwadata percobaan menunjukkan cocok dengan urutan keduapolinomial persamaan. Untuk semua model berkurang, P nilai< 0.05 wasobtained, menyiratkan model ini adalahsignifikan. Nilai presisi yang memadai adalah ukurangsignal (respon) h rasio kebisingan (deviasi). Rasio lebih besardari 4 yang diinginkan. Dalam studi ini, rasio ditemukan dikisaran 20.327-59.277, yang menunjukkan memadaisinyal dan, oleh karena itu, model ini penting untukproses pemisahan. Koefisien variasi adalah ukuranof reproducibility of the model and as a general rule a modelcan be considered reasonably reproducible if it is less than10%. The coefficient of variation for all the models wasfound to less than 10%. The results are shown in table2.Hence, the diagnostic plots, (a) the normal probability plotof the residuals and (b) the plot of the residuals versuspredicted values, were analyzed for response Rs2,3. Since theassumptions of normality and constant variance of theresiduals were found to be satisfied, the fitted model for theRs2,3was accepted. In table 2, the interaction term with thelargest absolute coefficients among the fitted models is AB(+0.37) of the tR3 model. The study reveals that changingthe fraction of MeOH from low to high results in a rapiddecline in the retention time at the low and high levels ofbuffer molarity. Further at the low level of factor B, anincrease in the buffer molarity results in a marginal decreasein the retention time. Therefore, when the MeOHconcentration is set at its lowest level, the bufferconcentration has to be at its highest level to shorten the runtime. This interaction is synergistic as it led to a decrease inrun time. In order to gain a better understanding of theresults, the predicted models are presented in figure 2 as theperturbation plot. For an optimization design, this graphshows how the response changes as each factor moves froma chosen reference point, with all other factors held constantat the reference value. A steep slope or curvature in a factorindicates that the response is sensitive to that factor. Hence,the plot shows that factor B mostly affected the analysis timetR3 followed by factor A and then C.Multi-criteria Decision MakingIn the present study, to optimize the three responseswith different targets, Derringer’s desirability function,was used.8The Derringer’s desirability function(D) isdefined as the geometric mean, weight, or otherwise, ofthe individual desirability functions. Desirability functiondi = Di(Yi)for each response separately (igoes from 1tothe number response say q) The expression that def