Eco-optimized RP-HPLC method for chiral tazobactam and piperacillin drugs: integration with Box–Behnken design and multi-sustainability tools

Abstract

Piperacillin-tazobactam (PIP-TAZ) is an effective antimicrobial agent, indicated primarily for severe infections resulting from susceptible organisms, and it has undisputed importance within antimicrobial therapy. PIP-TAZ combinations are widely used to treat severe infections caused by Gram-negative and Gram-positive bacteria, including those that produce β-lactamases. Our study involved the development of an environmentally friendly RP-HPLC method for the simultaneous determination of TAZ and PIP. In lieu of conducting numerous experiments, an efficient method to achieve the three major parameters of chromatography—pH, flow rate, and column temperature—was optimized with the help of a Box-Behnken Design (BBD) to minimize variability. This approach streamlines the experimental process and enhances results reliability. By carefully controlling these parameters, researchers can achieve more consistent and reproducible chromatographic analyses. This research addresses the impact that these variables have on such critical responses as retention times for TAZ and PIP and their resolution. The identification of optimal conditions was based on desirable and overlay plots. The samples were dissolved in ethanol, water, and tetrabutylammonium hydroxide in the mobile phase and separated through chromatography (58.5:40:1.5, v/v/v), and the pH was maintained at 3.5 with phosphoric acid. The method used UV detection at 230 nm, an injection volume of 20 μL, and a column temperature of 30°C. It was run at a flow rate of 1.0 mL/min. It used a Symmetry Shield RP18 column for 100-micron pore size, 5-micron particle size, and 4.6-mm by 250-mm dimensions. A comparison of carbon footprint reduction index (CaFRI), analytical greenness (AGREE), analytical eco-scales, green analytical procedure index (GAPI), modified GAPI (MoGAPI), and analytical method volume intensity (AMVI), which authenticates the accuracy, precision, and robustness of the method, was made in conjunction with a lesser environmental impact.

Supplementary Information

The online version contains supplementary material available at 10.1038/s41598-026-44942-2.

Introduction

The PIP/TAZ antimicrobial combination has very strong activity against Gram-negative bacteria, making it a preferred prophylactic treatment for liver transplanted patients (Fig. 1a, b). A high incidence of post-transplant infections is present in pediatric liver transplant patients. This necessitates adequate drug exposure. Due to its kidney-primarily excretion path, PIP/TAZ can cause renal workload and acute kidney injury if used excessively¹. Aside from these factors, numerous other factors can compromise renal function after transplantation, including postoperative hemodynamic instability, nephrotoxic immunosuppressive medications, and drug interactions. The physiological changes associated with pediatric liver transplant recipients may alter the pharmacokinetics of piperacillin/tazobactam, requiring individualized dosing². A number of studies have shown that dose adjustments are often required when pediatric patients undergo complex abdominal surgery or transplants, as well as adult liver transplant recipients suffering from Gram-negative infections, biliary tract infections, enterobacteria, and ventilator-associated pneumonia^3,4.

Fig. 1.

Chemical structures of (a) PIP and (b) TAZ with atom numbers.

Chromatographic practices related to environmental consciousness are sensational. One of the trends that has been on the increase when using ethanol compared to other solvents is that it is safe and sustainable. Evidence is mounting that similar results can be obtained using substitute reagents for conventional solvents^5,6.

Scientists around the world are paying particular attention to the development of ecologically friendly testing procedures and minimizing health risks. This is being brought about by the growing realization of the environmental and social consequences of conventional analytical practices. These processes often involve hazardous chemicals, large waste volumes, and energy use. Green analytical chemistry (GAC) has become a subdiscipline of laboratory science that incorporates environmental responsibility^7,8.

Besides designing and introducing effective, reliable, and environmentally friendly analytical processes, GAC also strives to better the environment. Elimination or reduction of toxic reagents, use of renewable resources, maximum use of energy, minimum use of waste, and creation of safe materials and solvents are all applied. An example of a GAC practice is the obligation to protect human health and the environment against the usual methods of analysis^9,10.

Moreover, GAC assists in the development of research and development on instrumentation, miniaturization and solvent-free or solvent-reduced technologies. In addition, it reduces sample handling and exposure through real-time analysis. With the world becoming more environmentally conscious and sustainability issues becoming more popular, green analytical methods, in addition to being a scientific given, are also a moral and regulatory given^11,12.

The number of analytical procedures for the simultaneous determination of TAZ and PIP by HPLC is limited in the literature^13–17. Current approaches focus on individual or formulation-specific quantification, without extensive validation or environmental concerns. These approaches too, are founded on traditional trial-and-error optimization methods and use organic solvents, which are highly harmful to the environment. Therefore, strong, environmentally friendly, and systematically optimized HPLC techniques for analyzing TAZs and PIPs are required. Analytical method innovation is thus imperative to guarantee precision and accuracy of analysis and compliance with green chemistry principles.

USP monographs specify standardized HPLC assays and sample preparation protocols for PIP and TAZ for injection. Using an acetonitrile–water mobile phase and a 210 nm detection wavelength. By contrast, our optimized method uses ethanol, water, and tetrabutylammonium hydroxide as the mobile phase, and a 230 nm detection wavelength. By optimizing both the mobile phase and the wavelength, we are able to attain greater performance and environmental sustainability¹⁸.

TAZ and PIP have been successfully determined chromatographically. In contrast to conventional methods, this technique offers a sustainable RP-HPLC process. Instead of trial and error optimization, to study the most significant chromatographic variables, such as pH, flow rate, and column temperature, we utilize BBD in a statistically robust and efficient manner. This design increases method reliability by minimizing experimental variation. The application of green mobile phases (ethanol, water, and tetra butyl ammonium hydroxide) (58.5:40:1.5, v/v/v) shows the application of green analytical practices. The pH was equilibrated to 3.5 using phosphoric acid. Before this development, methanol and acetonitrile were the most commonly used chemicals, which were less environmentally friendly. Additionally, 230 nm UV detection and a controlled column temperature of 30 °C have also been added to enhance our method’s resolution and reproducibility. The approach was found to be accurate, precise, and robust with a lesser environmental impact than AGREE, analytical Eco-scale, GAPI, MoGAPI, and AMVI tools. Also, the CaFRI validation measure focuses on how the method has a low impact on the environment, and this has not been done in earlier research. Instead of justifying linearity, precision, and accuracy as earlier studies did, we employ green chemistry and statistical optimization. This brings an additional dimension to the study. Despite the lack of relevant stability-suggesting capability, our technique proves to be more robust, accurate, and environmentally friendly, which is a significant addition to sustainable pharmaceutical analysis.

Experimental

Chemicals and reagents

Centrient Pharmaceuticals (Rotterdam, Netherlands) supplied TAZ and PIP. Among the HPLC and analytical grade materials used in this study were ethanol, deionized water, tetrabutylammonium hydroxide, and phosphoric acid. We obtained tetrabutylammonium hydroxide from Sigma-Aldrich (USA) and Spectrum Chemical (USA). We purchased ethanol and phosphoric acid from Thermo Scientific (USA). For chromatographic analysis, milli-Q water purification systems were used to prepare deionized water in-house. An intravenous injection vial of PIP-TAZ Sandoz 4.5 g powder was purchased from El Ezaby Pharmacy in Egypt.

Instruments

Using a Waters Alliance e2695 HPLC system developed in Milford, Massachusetts (USA) equipped with a built-in autosampler (quaternary pump, degasser, sample heater and cooler), samples were analyzed using a Waters 2489 UV–Vis detector. The data was acquired and processed using Empower 3 software (Waters Corporation). Symmetry Shield RP18 columns (Waters, 4.6 × 50 mm, 5 µm particle size, 100 Å pore size) with integrated heaters/coolers were used for the chromatographic separation. Autosamplers support injection volumes ranging from 0.1 to 100 µl (standard) and keep samples at a temperature of 4–40 °Celsius.

Analyses were performed in Design-Expert v13 (StatEase Inc. Minneapolis, USA). The same software was used to fit models, perform analysis of variance, plot response surfaces, generate diagnostics, and perform numerical optimization (desirability).

Diluent preparation

The solvent used to prepare the standard, samples, and system suitability solutions was identical in composition to the mobile phase. This was to avoid solvent mismatch, and provide consistent chromatographic performance. The diluent was ethanol, water and Tetrabutylammonium hydroxide in 58.5:40:1.5 (v/v/v) proportions. Phosphoric acid (analytical grade) was used to adjust the pH to 3.5 ± 0.02. The reagents were all HPLC or analytical grade (ethanol and phosphoric acid from Thermo Scientific; tetrabutylammonium hydroxide from Sigma-Aldrich).

Precaution steps: Precisely weigh ethanol, water and tetrabutylammonium hydroxide into labeled glassware with calibrated glassware. Shake it in a clean glass container. Adjust the pH using Phosphoric acid. Through this, diluent was prepared and filtered using a 0.45 µm nylon membrane filter. After this, the diluent was ultrasonically degassed for 15 min.

Storage stability: The prepared diluent was stored in amber glass bottles at room temperature (20–25 °C) and consumed within 48-h to avoid microbial growth or pH drift. The same diluent used in the mobile phase was selected to guarantee the solubility of analytes. It was also selected to reduce peak distortion and remove variability in retention time. There was no major difference in retention or efficiency. This proves that the method can withstand the mismatch between the diluent and the mobile phases.

Chromatographic system

As part of the chromatographic system, ethanol, water, tetrabutylammonium hydroxide, and their respective volumes were used as mobile phases. This mixture’s pH was adjusted to 3.5 by adding phosphoric acid to ensure optimal separation conditions. Aromatic and conjugated substances were detected by UV absorbance at 230 nm. Stability and reproducibility were improved by maintaining the column temperature at 30 °C. The flow rate was 1.0 mL/min to give a balance between resolution and time of analysis. Tetrabutylammonium hydroxide serves as an ion-pairing agent in reverse-phase chromatography to improve ionic analyte retention and separation. A Symmetry Shield RP18 column with the following specifications: 100 Å pore size, 5 µm particle size, and dimensions of 4.6 mm × 250 mm. During the sequence, the injection volume was set at at 20µL.

Standard stock and working solutions

Accurately weigh 20 mg of PIP and 2.5 mg of TAZ (maintaining an 8:1 ratio) into a 50 mL volumetric flask, dissolve in approximately 10 mL of ethanol with sonication, and dilute to volume in the mobile phase to obtain a stock solution of (PIP 400 µg/mL and TAZ 50 µg/mL). Preparing the working standard involves diluting 1 mL from stock solution and then transferring into a 10 mL volumetric flask. It also involves diluting the volume in mobile phase. This will yield PIP 40 µg/mL and TAZ 5 µg/mL, which fall within the validated range.

Sample solution

For the test solution, reconstitute the 4.5 g PIP/TAZ injection (4 g PIP + 0.5 g TAZ) in 50 mL to obtain PIP 80 mg/mL and TAZ 10 mg/mL, then dilute 0.25 mL from stock test solution to 500 mL and dilute with the mobile phase to achieve the same final concentration of PIP 40 µg/mL and TAZ 5 µg/mL for injection.

Establishing of calibration curves

The calibration curve for Piperacillin-Tazobactam was developed by preparing a series of standards with concentrations ranging from (1–10 µg/mL) for TAZ and (8–80 µg/mL) for PIP. Standard solutions were diluted serially using appropriate solvents from a stock solution. The validated chromatographic method was used to calculate the corresponding peak areas using each concentration. The calibration curve was plotted to ensure the relationship between concentration and peak area was linear. Following analysis by linear regression, a straight-line regression equation was obtained of the form y = mx + b, where y was the peel area and x was the concentration. Depending on the obtained curve, the drug may be quantitatively determined in pharmaceutical preparations since it has high linearity.

Method development strategy

There was a methodical approach to determining the final chromatographic parameters used in the simultaneous determination of TAZ and PIP. First, organic modifiers (methanol, acetonitrile, ethanol) were filtered to find a more green alternative. Ethanol has been chosen because it has a lower UV cut off, is more miscible with water and matches green analytical chemistry principles. The influence of pH on the shape and resolution of the peak was tested in the pH range of 2.5–6.0; a pH of 3.5 was adopted as the pH of sharp and symmetrical peaks and reproducible peaks. Systematic variation of column temperature (25 °C, 30 °C, and 40 °C), flow rate (0.8–1.2 ml/min) was done, and 30°C and 1.0 ml/min were found to be optimum to balance resolution and analysis time. A Symmetry Shield RP18 column (4.6 × 250 mm, 5 µm) was chosen to reduce silanol interactions and maximize peak symmetry. A Box-Behnken Design (BBD) was used to optimize these parameters by exploring the interactions between pH, flow rate, and temperature and all the effects on critical responses (resolution and retention times). Optimization was done using the response surface methodology. The ultimate method was chosen for obtaining a resolution of > 2.0, low tailing, reproducible retention times and adherence to the green principles of analytical chemistry. The mobile phase composition (ethanol: water: TBAOH, 58.5:40:1.5 v/v/v) was fixed based on initial screening experiments that showed the best peak shape, resolution and observation of green chemistry considerations; its predictable linear behavior and significance to keep the method green warranted its omission of the Box-Behnken Design.

Method validation

To gain reliability and reproducibility, the developed RP-HPLC method was calibrated according to ICH Q2(R2). The parameters measured include:

Linearity and Range Six concentration levels of piperacillin (8–80 µg/mL), and tazobactam (1–10 µg/mL) were plotted. Injection was repeated three times in each level. Acceptance criteria: correlation coefficients (r²) ≥ 0.999 and residuals are randomly distributed.

Accuracy Recovery experiments were done at 50, 100 and 150 percent of the assay concentration with pre-analyzed samples spiked with known quantities of standard. Acceptance range: 98–102% recovery for both analytes.

Precision Repeatability was measured by six replicate injections of a standard solution with a concentration of 100% assay concentration under the same conditions. Intermediate precision was performed on other days by analysts using the same equipment. Acceptance criterion: %RSD ≤ 2% for peak area and retention time.

Specificity Blank, placebo, and spiked samples were tested to verify the lack of interference in the retention time of piperacillin and tazobactam. There were no co-eluting peaks in the chromatograms.

Robustness Careful changes in chromatographic conditions were made, including pH (± 0.2 units), flow rate (± 0.1 mL/min), and column temperature (± 2 °C). Acceptance criterion: The parameters of system suitability did not exceed the limits (resolution ≥ 2.5, tailing factor ≤ 2.0).

LOD and LOQ:

Calculated based on the ICH formula:

S is the slope of the calibration curve, and σ is the standard deviation of the response.

System suitability:

Assessed prior to analysing the samples. This was based on: resolution ≥ 2.5 between critical peaks, tailing factor ≤ 2.0, theoretical plates ≥ 2000, and percent RSD ≤ 2% between replicate injections.

Results and discussion

Methods development and optimization

The initial tests were conducted in a range of organic solvent ratios with water, such as methanol and acetonitrile. However, these combinations yielded low resolution of the peak and tailing, especially with TAZ. Following a careful assessment of organic modifiers, ethanol was identified to possess a lower UV cut-off and better miscibility than water, so it was regarded as an alternative modifier. The most effective synthetic mobile phase was established between ethanol, water, and tetrabutylammonium hydroxide. They discovered that a ratio of 58.5:40:1.5 (v/v/v) was the most appropriate ratio to use to give the most appropriate balance between peak shape and retention time. Both filler ions showed a significant influence of pH on ionization in the mobile phase. The tests were conducted at a pH of 2.5–6.0. Peak broadening was significant at higher pH values and peak resolution decreased. A pH of 3.5 was established as optimal, which gave sharp and symmetrical peaks and increased reproducibility. During all trials conducted in the pH range of 2.5–6.0, the mobile phase composition (58.5:40:1.5, v/v/v) was maintained. Robustness was evaluated only by adjusting the pH of the aqueous phase. The column was tested at three temperatures: 25 °C, 30 °C, and 40 °C. The experiment was carried out at 30 °C to ensure minimal noise levels at baseline and the same retention time. The most effective flow rate was 1.0 mL/min and this gave optimal resolution and analysis time. The first tests were done at 210, 220, 230 and 254 nm for UV detection. It was established that the optimum wavelength of 230 nm was the most effective wavelength as both PIP and TAZ were absorbed at this wavelength well giving high sensitivity and low interference. A number of systematic trials were undertaken to establish the optimal resolution, maximum symmetry and sensitivity of the chromatographic technique. This was done by analyzing both PIP and TAZ. This paper relied on reverse-phase HPLC and a Symmetry Shield RP18 column (100 A, 5 um, 4.6 mm × 250 mm) was chosen to remove silanol interactions, which can interfere with moderate polar compounds, affecting the peak shape. The optimized method was adjusted and PIP and TAZ were well separated and had well-resolved peaks in less than eight minutes (Fig. 2). The method analyzed PIP-TAZ formulations successfully and was robust, reproducible, and suitable for daily quality control.

Fig. 2.

HPLC chromatograms of TAZ (5 μg/mL) and PIP (40 μg/mL) measured at 230 nm in (a) standard, (b) blank, and (c) sample solutions, with injection volume 20 μL.

Optimization of the RP-HPLC method using BBD

A screening study was conducted to identify the most influential parameters affecting chromatographic separation. This method was optimized through the use of a BBD combined with Response Surface Methodology (RSM). We evaluated three independent variables at three levels: pH, column temperature, flow rate, and their effects on resolution and retention time for the two target analytes^19–21. There were 15 runs of the BBD with 12 edge points and 3 centerpoint replicate. Using the BBD, three centerpoint replicates were used to determine pure error and curvature. Injections were made in duplicate for each experiment (edge and center), and the mean response was used to fit the model. The Method Validation section documents repeatability and instrument variability. During preliminary screening, the following factors were identified as critical: Factor A: pH of the mobile phase (Levels: 3.0, 3.5, and 4.0), Factor B: Flow rate (mL/min) (Levels: 0.8, 1.0, and 1.2), and Factor C: Column temperature (°C) (Levels: 25, 30, and 35). Three levels of each factor were studied, corresponding to low, center, and high values.

Temperature was chosen in the BBD because it has a more direct influence on the physicochemical properties of the mobile phase and analyte than does the organic phase composition. In particular, temperature affects mobile phase viscosity, which affects column backpressure and mass transfer kinetics. Consequently, the retention time and peak shape can vary significantly. Contact between Solute and Stationary Phase: The rising or falling temperature changes the solute and stationary phase equilibrium, which also influences selectivity and resolution compared to fine changes in organic composition within the studied range. Thermal stability and robustness of a method can be used to determine its robustness and stability under different laboratory conditions. These properties are critical to reproducibility and regulatory compliance. Although the phase composition in the organic phase is significant, its influence is relatively linear and predictable when compared to temperature. This can cause nonlinear behavior. The optimization model reflects this complicated interaction by focusing on temperature. The reason as to why resolution factor (RS) was selected over theoretical plate (N), height equivalent to a theoretical plate (HETP) or tailing factor, which was the case in this study, is simply because: Firstly, a major criterion of suitability of methods, particularly when dealing with multi-component systems, is the capability of separating analytes with adequate resolution, as is established in ICH Q2 and USP guidelines. Quantification can only be made if separation quality is reasonable and this can be directly measured by RS. As far as the robustness of the method is concerned, the theoretical plate (TP), the HETP, and the tailing factor are critical for the symmetry of the peaks and the efficiency of the column. They are, however, nowhere near enough to guarantee a baseline separation. RS may result in insufficient RS even in high-efficiency situations. RS is more imperative than specificity. Various Physicochemical Properties: Although the two drugs may be different physically, they may overlap partially under different conditions because of their retention behaviors. Bulking RS is robust to design possibilities. RS, being an integrated measure, incorporates efficiency, selectivity and retention into one measure. This is what makes it a more convenient optimization goal than separate parameters, such as theoretical plates or peak tailings.

The experiment had 15 runs, and retention time for TAZ and PIP was used as the main response variable. Table 1 summarizes the results of modelling with a second-order polynomial equation. Three-dimensional response surface plots were produced to visualize response surface interactions. Figures S1 to S3 present overlay plots with a desirable function for predicting the most appropriate conditions under which the method is to be used.

Table 1.

A design matrix showing 15 runs of resolution, TAZ retention time, and PIP retention time as a function of pH, flow rate, and column temperature.

		Factor 1	Factor 2	Factor 3	Response 1	Response 2	Response 3
Std	Run	A:pH	B:Flow Rate	C:Column temperature	Resolution	TAZ tR	PIP tR
6	1	3	0.8	30	2.08	4.67	7.12
4	2	4	0.8	30	2.06	4.11	6.88
2	3	3	1.2	30	2.12	4.56	7.05
15	4	4	1.2	30	2.06	4.15	6.91
5	5	3	1	25	4.53	6.04	7.45
1	6	4	1	25	4.67	4.54	7.32
12	7	3	1	35	4.59	4.22	6.98
11	8	4	1	35	4.6	4.64	6.55
9	9	3.5	0.8	25	4.44	4.82	7.3
8	10	3.5	1.2	25	4.48	4.8	7.28
14	11	3.5	0.8	35	4.51	4.82	6.6
13	12	3.5	1.2	35	4.45	4.7	6.5
10	13	3.5	1	30	2.02	3.886	6.688
7	14	3.5	1	30	1.97	4.12	6.85
3	15	3.5	1	30	1.99	4.14	6.87

With RSM, we maximize robust and efficient chromatographic procedures. BBD is a cost-efficient experimental design. Predictive models may be generated with the smallest number of experiments. BBD’s second-order model does not require drastic combinations of factors, which is especially beneficial for method development. The BBD approach does not involve corner points in the experiment space. This would reduce the predictive power for responses influenced by the extreme values of the independent variables. These drawbacks do not negate the BBD method as one of the most effective and viable methods of improving chromatographic conditions.

Discussion of statistical analyses findings

This research presented useful information regarding the performance of the model on each response variable according to statistical analysis of the chromatographic approach. The resolution quality was acceptable, as the mean value was 3.37 and the coefficient of determination was only 1.23%. This shows a high degree of accuracy. R², 0.9996, 0.9990, and 0.9948 indicate that the poly-model was closely related to the experiment data, implying high model fit and predictive accuracy (Table S1). In comparison to this, the TAZ retention time model exhibited moderate variation with a standard deviation value of 0.2721 and a mean of 4.55 min. Although the adjusted R² was 0.9008, the expected R² was −0.4404, which implies no adequate statistical power. Although the model works quite well in terms of fitting the available data, it is not very reliable in terms of predicting new results. It might require further improvement or expansion.

PIP retention time was well predicted by the model, with a standard deviation of 0.0800, a mean of 6.96 min, and a coefficient of determination of 1.15%. All three R² values 0.9741, 0.9275, and 0.8074 support the model’s robustness and predictive ability. In summary, the statistical analysis shows that resolution and PIP retention time models are highly reliable, while TAZ retention time model may require further optimization.

Factors influencing variables

Factors affecting resolution

This quadratic model evaluating the response variable “Resolution” shows a highly significant F-value of 1492.57, and a p-value less than 0.0001. Hence, the model effectively explains the variability in resolution across experimental designs. Although the pH, flow rate, and column temperature were linearly significant, their p-values were higher than 0.05, so they were statistically insignificant. A significant relationship between flow rate and pH, pH and temperature as well as flow rate and temperature did not exist between these three interactions. Rather, the quadratic terms yielded important insights: pH (A²) was highly significant (p 0.0001) and temperature (C²) was highly significant (p = 0.0043), indicating that these factors strongly affected resolution in a nonlinear manner. In this case, the lack of fit was not significant (p = 0.2116), indicating that the model does not suffer from systematic errors. Despite the minimal linear and interaction effects, the curvature effects of pH and temperature, particularly temperature, exert a significant influence on resolution (Table 2).

1

Table 2.

Analyses of ANOVA for significant models and factors affecting resolution.

Source	Sum of Squares	df	Mean Square	F-value	p-value
Model	23.22	9	2.58	1492.57	< 0.0001	Significant
A-pH	0.0006	1	0.0006	0.3544	0.5776
B-Flow Rate	0.0001	1	0.0001	0.0289	0.8716
C-Column	0.0001	1	0.0001	0.0651	0.8088
temperature
AB	0.0004	1	0.0004	0.2314	0.6508
AC	0.0042	1	0.0042	2.44	0.1787
BC	0.0025	1	0.0025	1.45	0.2829
A2	0.0423	1	0.0423	24.50	0.0043
B2	0.0015	1	0.0015	0.8905	0.3887
C2	23.02	1	23.02	13,320.99	<  0.0001
Residual	0.0086	5	0.0017
Lack of Fit	0.0074	3	0.0025	3.88	0.2116	Not significant
Pure Error	0.0013	2	0.0006
Cor Total	23.23	14

Using this equation, we can describe how the resolution is influenced by three factors: pH, Flow Rate, and Column Temperature). All factors are centered at the point (1.99), which is the predicted resolution. In all three variables, the linear coefficients (A, B, C) are small and positive, suggesting that individual changes in pH, flow rate, or temperature have a slight positive effect on resolution. AB, AC, and BC have negative interaction terms, indicating that combining these factors may decrease resolution slightly. There are four key insights revealed by quadratic terms:

Interestingly, the A² (pH squared) has a moderate positive trend (+ 0.1071), indicating that at certain pH levels, resolution is likely to improve. The increased flow rate has a small negative effect on point B² (Flow Rate squared), signaling the presence of a mild curvature and possibly an optimum point. The temperature-squared coefficient (C²) is very positive (+ 2.50), indicating that temperature impacts resolution in a nonlinear way. The results suggest that resolution increases significantly at certain temperatures, which is an important factor in optimizing.

The two factors that affect tazobactam retention time the most or a combination of the two are pH and temperature. This parameter is to be optimized to maximize retention characteristics because it has a strong quadratic effect (Figs. 3, 4 and 5).

Fig. 3.

Plots of contours and 3D responses showing how flow rate and pH influence the responses to resolution (a, b), TAZ retention time (c, d), and PIP retention time (e, f).

Fig. 4.

Plots of contours and 3D responses showing how column temperature and pH influence the responses to resolution (a, b), TAZ retention time (c, d), and PIP retention time (e, f).

Fig. 5.

Plots of contours and 3D responses showing how flow rate and column temperature influence the responses to resolution (a, b), TAZ retention time (c, d), and PIP retention time (e, f).

Factors affecting TAZ retention time

As shown in the ANOVA results (Table 3) for Response 2: Tazobactam Retention Time (TAZ tR), the quadratic model appears to be statistically significant as indicated by an F-value of 5.04 and a p-value of 0.0448. The reported retention time of Tazobactam across experimental conditions can be explained by the model adequately.

Table 3.

Analyses of ANOVA for significant models and factors affecting TAZ retention time.

Source	Sum of Squares	df	Mean Square	F-value	p-value
Model	3.36	9	0.3733	5.04	0.0448	Significant
A-pH	0.5253	1	0.5253	7.10	0.0447
B-Flow Rate	0.0055	1	0.0055	0.0745	0.7958
C-Column	0.4141	1	0.4141	5.59	0.0643
temperature
AB	0.0056	1	0.0056	0.0760	0.7938
AC	0.9216	1	0.9216	12.45	0.0168
BC	0.0025	1	0.0025	0.0338	0.8614
A2	0.1468	1	0.1468	1.98	0.2180
B2	0.0572	1	0.0572	0.7722	0.4197
C2	1.38	1	1.38	18.68	0.0076
Residual	0.3701	5	0.0740
Lack of Fit	0.3302	3	0.1101	5.52	0.1573	Not significant
Pure Error	0.0399	2	0.0199
Cor Total	3.73	14

There was a significant effect of pH (A) among the individual factors (F = 7.10, p = 0.0447), showing that pH has a meaningful impact on Tazobactam retention. There is a significant effect of column temperature (C) as well (F = 5.59, p = 0.0643), which is close to the 0.05 threshold but not significant enough to indicate that it might have an important influence. pH and temperature (AC) interact significantly (F = 12.45, p = 0.00168), indicating that both factors affect retention time jointly.

The non-linearity of this correlation is very high, as the quadratic temperature term (C2) is very significant (F = 18.68, p = 0.0076). Instead, none of the pH (A2), flow rate (B2) or any other interaction terms (AB, BC) are significant, which argues that the curvature or interactions are small. Although the model is not perfectly fitted (p = 0.1573), it fits the data without systematic error. The discussion has found that pH and column temperature (and their combination and the non-linear influence of temperature) have the most significant effect on tazobactam retention time, and flow rate has a rather insignificant effect.

In order to calculate the Tazobactam retention time (TAZ tR), the following regression equation is used:

2

It is possible to model tazobactam retention time as a function of pH, flow rate and column temperature (A, B, C) based on this equation. The intercept (4.05) represents the projected retention time (4.05). On a linear basis, when compared to flow rate and temperature, pH (0.2562) and temperature (0.2755) influence retention time the most. Through the interaction terms, it is clear that the interaction effects that pH and temperature (AC) have are very beneficial in terms of retention time in the form of + 0.4800. We have minor interactions with BC and AB. Quadratic terms show U-shaped relationships:

C² (square of temperature) has the strongest effect (+ 0.6119), indicating that retention time is longer at low and high temperatures. But the role of pH squared and flow rate squared in curvature is minor (Figs. 3, 4 and 5).

Influencing factors of PIP retention time

ANOVA findings (Table 4) were applied to the analysis of Response 3: PIP Retention Time (PIP tR) and give an idea of the variables that influence this response.

Table 4.

Analyses of ANOVA for significant models and factors affecting PIP retention time.

Source	Sum of Squares	df	Mean Square	F-value	p-value
Model	1.20	9	0.1336	20.89	0.0019	Significant
A-pH	0.1105	1	0.1105	17.27	0.0089
B-Flow Rate	0.0032	1	0.0032	0.5004	0.5109
C-Column	0.9248	1	0.9248	144.62	<
temperature					0.0001
AB	0.0025	1	0.0025	0.3910	0.5592
AC	0.0225	1	0.0225	3.52	0.1195
BC	0.0016	1	0.0016	0.2502	0.6382
A2	0.1082	1	0.1082	16.92	0.0092
B2	0.0010	1	0.0010	0.1509	0.7137
C2	0.0378	1	0.0378	5.91	0.0593
Residual	0.0320	5	0.0064
Lack of Fit	0.0121	3	0.0040	0.4032	0.7686	Not significant
Pure Error	0.0199	2	0.0100
Cor Total	1.23	14

This model is statistically significant with a value of 20.89 and a p-value of 0.0019, which shows that it can be relied upon to explain piperacillin retention time variation. pH (A) and column temperature (C) are the two most crucial factors. PIP retention is highly influenced by pH (F = 17.27, p = 0.0089), indicating that pH changes are significant. The effect was more prominent when column temperature was considered (F = 144.62, p = 0.0001).

pH (A²) showed a nonlinear correlation with retention time (F = 16.92, p = 0.0092), and also with the square of it (A²). The quadratic term for temperature (C²) could be relevant to more observations (F = 5.91, p = 0.0593). On the contrary, flow rate (B) and terms related to flow rate (B, B², and interactions AB and BC) do not have any significant influence on piperacillin retention time in encountered ranges.

The model had no systematic errors (p = 0.7686), which proves it has no systematic errors. On the whole, the analysis indicates that pH and column temperature are very crucial in determining piperacillin retention time. However, flow rate does not have much influence.

3

The three factors that influence piperacillin retention time are pH, flow rate, and column temperature. The intercept (6.80) depicts all of these factors. pH, flow rate, and temperature cause a decrease in retention time, with the biggest influence being temperature (C), meaning that increased temperatures cause piperacillin to be retained significantly. A low interaction value (AB, AC, BC) implies that there are only slight interactions between the factors. The AC negative sign (−0.0750) shows that the higher the pH and temperature, the shorter the retention time. Due to quadratic terms, the response surface is curvy.

A² (squared of pH) and C² (squared of temperature) have a U-shaped relationship, indicating that piperacillin’s retention time is raised at both ends of pH and temperature.

The effect of flow rate squared (B²) is slightly positive, which means slight curvature with insignificant practical implications (Figs. 3, 4 and 5).

Column temperature has the strongest linear and quadratic impact on piperacillin retention time. This is followed by pH and flow rate, with pH and temperature disturbed relative to each other to produce optimal retention.

PIP retention time (PIP tR) can be modeled by the following regression equation:

Sustainability tools

CaFRI

By evaluating the environmental impact of defined analytical laboratory procedures, the CaFRI focuses on their environmental impacts. The assessment will consider numerous critical elements such as energy use, carbon footprint prevention strategies, storage of samples, transportation, human involvement, waste disposal, resource conservation programs, and chemical use. Out of the analytical methods evaluated in the CaFRI assessment, 100 is the highest score. The tool (https://bit.ly/CaFRI) is an intuitive user interface that is free to use online. Green chemistry principles say every region on the footprint diagram should depict a particular criterion. Red indicates a poor rating, yellow indicates an average rating, and green represents a favorable rating. Based on the questionnaire findings, it is possible to get a total score of 100 for a fully environmentally friendly procedure that considers its carbon footprint^22,23.

Figure S4a presents the results of applying CaFRI to the proposed HPLC- UV detector method in addition to the official and reported ones. A cumulative score of 88 from CaFRI indicates that the proposed method makes significant efforts to reduce the carbon footprint (Fig. 6a).

Fig. 6.

Sustainable tools for evaluation HPLC method (a) CaFRI, (b) AGREE, (c) GAPI, and (d) MoGAPI.

AGREE

Analytical methods are evaluated for environmental sustainability through the AGREE tool. Among the 12 criteria evaluated are those aligned with GAC, such as minimization of sample treatment, waste reduction, avoiding toxic reagents, and energy efficiency. Based on each criterion, the method’s greenness is rated from 0 (least green) to 1 (most green). According to the AGREE tool analysis of the proposed HPLC method, it is observed that the method’s greenness is quite high. The average score was approximately 0.72 for the 12 criteria (Fig. 6b). The technique scored well for minimal sample size, energy savings, lack of derivatization, and the use of toxic reagents. These areas were all scored just below or below the maximum mark. As a result of these points, the method is both efficient and safe for the operator and the environment as a whole. Nevertheless, some weaknesses were recognized, such as zero in in-situ measurement, since this implies that the procedure needs to be analyzed in the laboratory and not directly in the field. Production of analytical waste and low consumption of renewable reagents were also rated moderately, and it is possible to improve the situation. These limitations are due to the use of phosphoric acid and conventional solvents in the mobile phase. To make the method more green, they might include minimizing wastes by means of shorter run times/smaller column sizes, the use of bio-based or less toxic solvents, and additional automation/miniaturization (Figure S4b). In general, although the technique complies sufficiently with some GAC concepts, specific modifications would make it more sustainable. This is without negatively affecting analytical standards²⁴.

GAPI

GAPI is an inclusive tool used to analyze the effects on the environment of analytical procedures. It addresses the whole analytical process—collection and preparation of samples, Instrumentation, and waste management according to GAC. GAPI shows the ratio of greenness of the stages in a pentagonal diagram and with colors. Green is environmentally friendly and is calculated as 1 point, yellow is moderate and is calculated as 2 points and red is something that needs to be improved and is assessed as 3 points. There is no difficulty in determining which steps of the method are sustainable with this visual approach, and which one needs to be optimized²⁵. The proposed HPLC method has a scatter plot on the GAPI pictogram with green, yellow, and red areas (Figure S4c). This indicates both the advantages and disadvantages of using it in the environment Green areas must refer to a powerful chromatographic method, no derivatization, and safe work conditions. Nevertheless, the red color denotes that there are major environmental issues, especially when it comes to the preparation of the sample and the use of solvents. The use of traditional organic solvents and phosphoric acid presents a high environmental impact. Yellow areas indicate medium levels of sustainability associated with energy use and generation of waste in the analysis. On the whole, the pictogram shows that even though the process is efficient and corresponds to green principles, the reagent choice, waste minimization, and even miniaturization could be improved to make the process more green (Fig. 6c).

MoGAPI

An enhanced version of the original GAPI tool, MoGAPI is a tool used to quantify and offer a more detailed analysis of the environmental impact of analytical methods. It assesses the various steps of analytical stages, sample collection, preparation, the use of reagents and solvents, instrumentation, waste treatment and quantification. This is done with a structured questionnaire and a color-coded pictogram. They have been rated and indicated in green (environmentally friendly), yellow (moderate impact), or red (high impact) which provides a clear visual overview of the method’s sustainability²⁶. MoGAPI also adds a numerical value (85/100) to measure overall greenness, so they can be easily compared. The MoGAPI analysis of the suggested HPLC technique has an excellent sustainability profile with an overall score of 85 (Fig. 6d). This shows high compliance with environmentally friendly analysis principles. Most pictograms are green. The result implies the following positive characteristics, which include easy sample preparation, utilization of green solvents, low reagent levels (less than 10 mL), and non-toxic instrumentation with low power consumption (≤ 0.1 kWh/sample). Moreover, waste management was also effective since waste levels were managed, and techniques like degradation and passivation were employed. Nevertheless, there are some yellow spots indicating a moderate concern for the environment These spots could be linked to off-line sampling and the lack of in-situ analysis which restricts greenness in real time. There were no red zones, which proves that the method does not imply significant risks, including highly toxic reagents or excessive waste. MOGAPI results indicate that the approach is very sustainable (Figure S4d).

Analytical eco-scale

The Analytical Eco-Scale is used to determine whether analytical methods are environmentally friendly. Base scores are 100, and penalty points are deducted for factors that negatively impact sustainability, such as hazardous reagents, high energy consumption, waste production, and occupational hazards. According to the final score, the method is green: ≥ 75 = Outstanding green analysis, 50–74 = Passable green analysis, and < 50 = Poor green analysis²⁷. The HPLC method proposed was evaluated in the Eco-Scale that would have 13 penalty points and the final score of 87 would result in the classification of the method as an excellent green analysis (Table 5). Ethanol, phosphoric acid, tetrabutylammonium hydroxide, energy consumption used to operate the HPLC, ultrasonic use and generation of waste were the key contributors to the penalty points (4-points, 2-points, 1-point, and 3-points respectively). The level of purified water and occupational hazards was zero, which means that the impact on the environment was low in these regions. In general, the approach is highly adherent to green chemistry principles with few penalties associated with the toxicity of solvents and wastes. Additional gains might be gained by substituting ethanol with a greener one and minimizing waste to get a better sustainability rating.

Table 5.

An analysis of the proposed method to estimate PIP and TAZ using the Eco-scale.

Proposed Method	Analytical Eco- scale	Penalty points
Reagents	Purified water	0
	Phosphoric acid	2
	Ethanol	4
	Tetrabutylammonium hydroxide	2
Instruments	Energy for HPLC < 1.5 kWh /sample	1
	Occupational hazard	0
	Ultrasonic	1
	Waste: (1–10 mL = 3)	3
	Total penalty points	13
	Eco-Scale total score	87

Protocol for analytical eco-scale calculations

According to Gałuszka et al. ²⁷, Eco-Scale scores are based on subjective assessment. For example, recent literature assigns penalty points for reagent hazards, energy consumption, occupational hazards, waste generation, and auxiliary processes for each method starting at 100 points.

Penalty assignment:

Reagents are evaluated according to their hazardous nature (e.g., purified water is 0, ethanol is 4, phosphoric acid is 2, TBAOH is 2 points).

Instrumentation: HPLC energy of 1.5 kWh/sample = 1 point, ultrasonic energy = 1 point, occupational hazard = 0.

A score of 3 points is given for the total method waste (1–10 mL classified as moderate).

The penalty values are based on the standardized GAC literature and are detailed in the supplement.

Using the following formula:

Eco-Scale Score = 100 − Total Penalty Points.

According to the proposed method, the total penalty points (13 points) will be 13; the Eco-Scale score will be 87 (Table 5).

AMVI

The AMVI tool is supposed to evaluate the intensity of solvent use in an analytical procedure and give a quantitative value of how efficient the process is. It sums up all the solvent volumes consumed at all levels of HPLC operation and sample preparation and assigns it a score. This can be employed to find more sustainable and cost-efficient approaches by reducing the amount of solvent used²⁸. This is one of the fundamental principles of green analytical chemistry In the case of the proposed HPLC procedure, the AMVI analysis indicates that the total amount of solvent consumed is 392 mL whereby 192 mL is consumed in the HPLC runs and 200 mL is used during sample preparation. The AMVI score obtained is 196, which indicates moderate volume intensity. The breakdown reveals that 49 percent of all solvents and 51 percent of all solvents are used in sample preparation and HPLC, respectively (Table 6). This information may be understood to show that the two steps have an almost equal distribution. Although the process seems moderately efficient, the amount of solvents used is quite high. This has been caused by injections at large preparation volumes. To increase AMVI scores, it can be implemented to decrease run-time, maximize injection numbers and use as little preparation time as possible. These changes would enhance the method’s sustainability and the analysis’s performance.

Table 6.

The AMVI score of the developed approach for measuring PIP and TAZ.

Criteria	Proposed Method
Solvent Consumption HPLC (mL) =	192
Flow rate	1.0
Run time	8
Number of injections for 1 complete analysis	12
Number of analytes	2
Solvent Consumption Sample Prep (mL) =	200
Standard Prep volume (mL)	50
Number of Std. Preps	1
Sample Preps volume (mL)	50
Number of Sample Preps	2
System Suit volume (mL)	50
Number of System Suit Preps	1
Total Method Solvent Consumption =	392
Analytical Method Volume Intensity	196
% Consumption HPLC	49
% Consumption Preparations	51

Protocol for AMVI calculation

Based on the quantities of solvents used in each run (preparation, injection, and system suitability), AMVI sums the total volumes of solvents used. The following is based on Hartman et al.²⁸:

Solvent Consumption in HPLC: Calculated as flow rate (mL/min) × run time (min) × number of injections.

Sample and Standard Prep: Sum of all volumes for sample, standard, and system suitability preparations.

Calculation:

AMVI Score

Consumption percentages calculated as

% Consumption: 49% HPLC, 51% preparation.

Reconciliation of waste A complete batch of AMVI solvent consumption is calculated by adding up all sample, preparation, and cleaning steps for each batch, while Eco-Scale waste (1–10 mL) refers to waste generated per analytical cycle (Table 6).

A comparative analysis of greenness is conducted

There is a clear advantage of the suggested HPLC method over the reported ones^13–17 either in terms of ability to analyze or in terms of environmental friendliness. Compared to the reported methods, the proposed method involves systematic Box-Behnken Design (BBD), ensuring robustness and efficiency of parameter selection over trial and error.

The ethanol: water: TBAH (58.5:40:1.5, pH 3.5) mobile phase in the proposed method is cleaner than reported methods, where many of them require methanol or acetone as well as phosphate buffers, which are not so environmentally friendly. The proposed method uses a consistent flow rate (1.0 mL/min) and a detection wavelength (230 nm) with the standard practices, although temperature control (30 °C) is added to enhance the reproducibility of the method. The proposed HPLC method is greener than the reported methods since it requires minimal solvent and energy, simplifies the sample handling process, and does not use harsh reagents (Table S2). It has an Eco-Scale of 87 (excellent) and an AMVI of 196, with solvent use of 392 mL per full analysis (49% HPLC, 51% preparation), which is lower than reported methods, with lower Eco-Scale scores (70–80) and higher AMVI values (260–496), indicating that it uses a greater intensity of solvent^13–17. The AGREE/GAPI/MoGAPI plots of our approach show more green areas because (i) non-derivatized preparation was used; (ii) operating conditions were mild (30°C, 1.0 mL min⁻¹, 8-min run, UV at 230 nm); (iii) the amount of hazardous reagents incorporated was small (ethanol in low quantities and phosphoric acid in low amounts); (iv) the quantity of waste was moderate (1–10 mL) and the energy consumed per sample was low (Table S3). Conversely, the bioanalytical technique of human plasma^15,17, inevitably leads to low greenness since matrix cleanup (protein precipitation, LLE/SPE) introduces several preparation steps, increased solvent volumes, and more operator contact—penalized by AGREE/GAPI and indicated by high AM VI. The other method, which is reported¹⁶, has had a lower score in Eco-Scale since it intentionally incorporates high acid/alkali/oxidants and other stress measures to produce more hazardous waste and extra injections/runs, which have a penalty point in terms of reagent hazard, waste volume, and energy taken. To sum up, the proposed method is evidently superior to the reported ones because of its simplified structure and the principle of the green analytical approach. It does not require complicated sample treatment and extraction procedures, as required in the plasma-based methods, resulting in less reagent consumption, solvent usage, and exposure to the operators. This ease, coupled with less harsh operating conditions and use of less risky reagents, leads to increased AGREE, GAPI, and MoGAPI scores. The approach also has an Eco-Scale score of 87, versus the 70–80 of the reported approaches, since it does not entail forced degradation experiments that use strong chemicals and produce vast amounts of toxic chemicals. The solvent intensity is also significantly lower in its AMVI score of 196 as compared to 260–496 when other methods are used. Conversely, plasma analysis methods receive lower scores because they require several cleanup procedures and voluminous solvents, whereas degradation-oriented methods receive penalties because of harsh stress environments and extra runs. All of that is why the offered technique is greener, safer, and more resource-efficient than the techniques that were reported earlier. Table S4 is a comparative analysis of the environmental impact of the proposed RP-HPLC procedure and the USP procedure¹⁸ with PIP and TAZ, and it shows significant differences. The suggested approach scores higher (87) on the Eco-Scale than the USP method (80), which is more green. This can be attributed to the fact that ethanol was used as the organic modifier in lieu of acetonitrile. This has a higher hazard classification and a penalty of 6 points in the USP approach than 4 points with ethanol in the proposed approach. Moreover, the suggested approach adds tetrabutylammonium hydroxide (TBAOH (with a slight penalty (2 points)) to it, but this does not compensate for the benefit of not using acetonitrile.

The amount of waste generated by each run is also vastly different. The proposed method generates 192 mL/run, with 3 penalty points assigned, and the USP method uses 400 mL/run, with 5 penalty points awarded. This is the difference in the higher demand for solvent in the method of USP, which is consistent with the higher level of its use in the method of HPLC (55%) than the proposed method (49%).

According to AMVI approaches, the two approaches have identical solvent intensity (196 mL/analyte in the proposed method and 200 mL/analyte in the USP method¹⁸) since the batch assumptions are the same. Nevertheless, the composition of the solvents is very significant: the use of acetanilide as the solvent in the USP method is associated with an increased environmental impact for the same volume.

In general, the proposed technique proves to be more green since it minimizes the use of hazardous solvents and minimizes fines from waste. In addition, its analytical performance is similar. However, the two approaches demonstrate a fairly high level of solvent use. This highlights the necessity of additional optimization options like decreasing the run time or implementing gradient elution using less eco-friendly solvents.

Method validation

According to ICH standards, development and validation of the HPLC procedure was conducted.²⁹

Linearity and range

Based on the peak area data, calibration curves for PIP (8–80 µg/mL) and TAZ (1–10 µg/mL) were established. Serial dilutions of a mixed stock solution (PIP 800 µg/mL, TAZ 100 µg/mL, 8:1) prepared six paired levels in 50 mL volumetric flasks to ensure simultaneous coverage of all validated levels. Regression analysis yielded highly linear responses for both analytes: PIP (slope = 493,761.26, intercept = –88,453.15, r² = 0.99997) and TAZ (slope = 115,039.14, intercept = 345.43, r² = 0.99999), see Table 7 and Figure S5. According to back-calculated concentrations, PIP ranged between 98.95% and 101.05% of nominal concentration, while TAZ ranged between 99.56% and 100.16% (Table S5). An analysis of residual plots confirmed that the distribution was random around zero, indicating there were no systematic biases. These findings indicate robust linearity of PIP and TAZ quantification accuracy, and precision over the required range.

Table 7.

An analysis of the regression for TAZ and PIP using the HPLC technique.

Parameter	Drug TAZ	PIP
Wavelength	230 nm	230 nm
Range (ug/mL)	1–10	8–80
Coefficients of determination (R2)	0.99999	0.99997
Slope (b)	115,039.14	493,761.26
RSD of the slope (Sb%)	0.3	0.9
Intercept (a)	345.43	-88,453.15
Sa	294.71	43,085.82
Sb	97.34	1778.85
Sy/x	694.02	101,461.88

^aLimit of detection (3.3 × o /Slope) and a limit of quantitation (10 × o /Slope).

LOD and LOQ

The two values were calculated using ICH-advocated equations to calculate the LOD and LOQ of the calibration curves:

The calculated LOD and LOQ indicated that PIP had LOD and LOQ of 0.29 and 0.87 µg/mL, respectively. TAZ had LOD and LOQ of 0.008 µg/mL and 0.26 µg/mL.

Under this procedure, trace-level analysis of pharmaceutical formulations is achievable with adequate sensitivity to both analytes (Table 7).

Precision

Repeatability

Six replicate injections of a 100% concentration of PIP 40 µg/mL and TAZ 5 µg/mL were compared under identical operating conditions and over a short period of time. This was done to determine repeatability, which is an essential element of method precision.

It was also clear that the results could be repeated, and the values of the percentage of the standard deviation of the peak area, retention time, and theoretical plate count were all within acceptable limits (usually < 2%). Particularly: PIP retention time and 0.9 percent RSD from the peak area. TAZ had 0.6% and 0.8% RSD for peak area and retention time, respectively. On the basis of these low RSD values, it can be stated that the method can be used to give consistent and reliable results even in challenging circumstances (Table 7).

Reproducibility

Reproducibility was also considered to establish uniformity of the process under different circumstances, including days, analysts, and instruments. Two independent analysts were used to analyze the standard solutions of PIP and TAZ utilizing the same chromatographic conditions at the same concentrations. This was done on two different days. Excellent reproducibility results were obtained and the RSDs of the peak area and retention time were within acceptable limits (they are usually less than 2%). Therefore, it can be used in quality control laboratories, as it is consistent and reliable in various operational environments (Table S6).

Accuracy and recovery

PIP and TAZ recovery studies were done using the standard addition technique at three concentration levels (50, 100 and 150 percent of the assay concentration) as a measure of the method’s performance. The mixtures were spiked with pre-analyzed samples spiked with standard solutions and then analyzed using the validated RP-HPLC method. To obtain the percentage recovery, the measured concentration was used to compare it against the known amount added. The results were very accurate, with a mean recovery within the acceptable range of 98–102% (Table 8).

Table 8.

Testing of the proposed HPLC method for accuracy and recovery.

Parameters	HPLC
Test level %	Nominal Concentration (ug/mL)	TAZ amount found (ug/mL)	TAZ Recovery (%)	Nominal concentration (ug/mL)	PIP amount found (ug/mL)	PIP Recovery (%)
50	2.5	2.483	99.30	20	20.250	101.25
		2.495	99.80		20.168	100.84
		2.473	98.90		19.978	99.89
		2.515	100.58		20.026	100.13
		2.511	100.45		20.338	101.69
		2.486	99.44		20.112	100.56
Mean ± RSD		2.49 ± 0.66	99.75% ± 0.66		20.15 ± 0.67	100.73% ± 0.67
100	5	5.013	100.26	40	39.344	98.36
		4.973	99.46		39.832	99.58
		5.072	101.43		40.184	100.46
		4.988	99.76		40.584	101.46
		5.014	100.27		39.936	99.84
		5.009	100.18		39.904	99.76
Mean ± RSD		5.01 ± 0.67	100.23% ± 0.67		39.69 ± 1.03	99.91% ± 1.03
150	7.5	7.459	99.45	60	60.096	100.16
		7.560	100.80		60.978	101.63
		7.581	101.08		59.922	99.87
		7.565	100.86		60.282	100.47
		7.559	100.79		59.802	99.67
		7.467	99.56		60.288	100.48
Mean ± RSD		7.53 ± 0.72	100.42% ± 0.72		60.23 ± 0.69	100.38% ± 0.69

Robustness

By adjusting the chromatography conditions, we were able to determine whether the technique was sound. Minor adjustments to the flow rate (± 0.1 mL/min), column temperature (± 2.0 °C) and mobile phase pH (± 0.22 units) were experimented with.

Due to these differences, system suitability parameters—retention time, resolution, and peak areas-did not vary significantly. In all cases, the RSD value was less than 2, which indicates that the technique is strong enough and gives consistent results in regular laboratories. Even though small changes in operations might be observed, these findings prove the stability and reliability of the method (Table S6).

Stability of standard solutions

The experimental results show that the standard solution is of high quality. The results of the solution prepared recently at room temperature and in the refrigerator were excellent compared to the solution prepared earlier. The standard that was previously stored is also well matched with the one that was recently made, with a standard deviation of less than 2.0. According to these results, the standard solution seems very stable and reliable and therefore a good alternative in any analysis (Table S6).

System suitability

System suitability testing (SST) is necessary to validate analytical methods, particularly chromatography. It is only after an analytical system is properly run and willing to give reliable results that reliable and reproducible analyses can be done. It is imperative to note that when assessing SST, several of its significant metrics are compared. These metrics include retention time, theoretical plates, tailing factor, resolution, and repeatability. An excellent retention time of 3.889 and 6.683 min, 3863 and 4769 theoretical plates (indicating excellent column efficiency), a tailing factor of 1.3 and 1.1 (indicating symmetrical peaks) and a percentage of reasonable standard deviation of 0.8 and 0.7% with six injections can be used to illustrate a high level of repeatability of TAZ and PIP. To have sufficient separation, a resolution of 2.9 between the critical peaks is sufficient. Due to the fact that all these values are within accepted limits (Table 9), it will be appropriate to analyze the system in this regard.

Table 9.

Overview of HPLC system suitability items

Suitability parameter	TAZ	PIP
% RSD tR	0.8	0.7
Resolution (NLT 2.0)	N/A	11.7
Purity angle	0.077	0.086
Purity threshold	0.273	0.280
Theoretical plates (NLT 2000)	3863	4769
Tailing factor (NMT 2.0)	1.3	1.1

Specificity

Blank samples, a placebo, and individual standards and spiked samples were injected to ensure the method could measure PIP and TAZ without interfering with the method. Specificity was achieved by the analysis of the analytes’ retention times, where no co-eluting peaks were obtained (Figure S6).

Selectivity

In analysis, selectivity is the method’s ability to accurately measure analytes despite impurities, degradation products or matrix elements. Selective procedures are used so that the signal obtained is only due to the analyte and not another substance.

As part of the selection process, blank samples, spiked samples, and samples containing potential interfering substances are analyzed (Figure S6). Detection wavelengths and retention times are used to determine whether the method is selective.

Assay of dosage form

An assay for a dose form such as Piperacillin-Tazobactam involves determining the level of active pharmaceutical ingredients in the reconstituted sample in order to verify that the labeled claim has been met. Piperacillin and tazobactam make up 4.0 g of the vial, which contains 4.5 g total. Analyses are conducted using HPLC, which injects a standard solution and a sample solution simultaneously. According to the calculated assay values, each component’s concentration was 102.3% and 97.6% when compared to the standard. As long as the values are within the acceptable range (typically 90–110%), the dosage form is considered to be of high quality (Table S7).

In use stability study results

Based on the in-use stability study findings, Piprataz 4.5 gm maintained its potency within acceptable limits. Initially, the assay results were in agreement with the claim on the label, and subsequent measurements at defined intervals showed that there had been no significant degradation. After reconstitution, the product was stable at room temperature of not more than 25 °C for 24 h, and at 2–8 °C for 48 h (Figure S7). The results confirmed that the formulation was chemically stable within the specified range of 90–110% of the nominal concentration. Storage and handling conditions were recommended to ensure chemical stability (Table S8).

Conclusion

RP-HPLC with an eco-optimized method for the concurrent measurement of TAZ and PIP has successfully integrated advanced experimental design with GAC principles. The BBD enabled the method’s optimization with minimal experimental runs, focusing on pH, flow rate, and column temperature, which are critical chromatographic parameters. An ethanol-based mobile phase was used in combination with water and tetrabutylammonium hydroxide in order to ensure effective separation and resolution of analytes and to minimize environmental impact. CaFRI demonstrated excellent performance for evaluating the method. The overall greenness evaluation of the analyzed HPLC methodology with the help of AGREE, GAPI, MoGAPI, Analytical Eco-Scale, and AMVI software proves that the method is predominantly sustainable and corresponds to GAC. Both AGREE and GAPI assessments have high scores in terms of energy efficiency, safety, and the avoidance of toxic reagents. However, they have slight weaknesses in in-situ analyses and renewable reagents. MoGAPI validates greenness with an 85/100 score, safe instrumentation, and appropriate waste treatment. Eco-Scale of 87 shows the method as fine green analysis, in which solvent choice and waste volume are not penalized. AMVI shows a medium solvent usage (392 mL), which implies that some run time and preparation procedure optimization is possible. Overall, the procedure is eco-friendly and safe in terms of its operation, and solvent reduction, waste minimization, and the use of greener reagents could be improved. An analysis of this study demonstrates the potential of combining statistical optimization tools with sustainability metrics for developing high-performance, environmentally responsible analytical methods to monitor antimicrobial formulations.

Study limitations and future work

This RP-HPLC method was validated for linearity, precision, accuracy, and suitability for routine quality control of multiple marketed products and dosage forms. Nevertheless, there are some limitations. Specifically, in this phase, forced-degradation studies were not performed to confirm stability-indicating capabilities, comprehensive impurity profiling in accordance with ICH Q3A/Q3B, or stability testing under clinical preparation and storage conditions. All of these aspects are essential for the full lifecycle applicability of a product, including stability programs and regulatory submissions. It is therefore planned to conduct systemic stress tests (acid–base, oxidative, thermal, and photolytic) as well as quantify peak purity and mass balance and characterize impurities in detail. With these additional studies, the method’s robustness will be further improved and its applicability for stability studies, manufacturing support, and extended QC use cases will be confirmed. Even though the current study utilized a Box-Behnken design to determine how pH, flow rate, and column temperature affect chromatographic performance, there were a few limitations that should be resolved. The optimization was performed under a small set of operational conditions, and curvature was dominant at column temperature. Most terms of interaction and flow rate showed negligible effects, which restricted true multi-factor optimization. This means that the results can only be seen as robust but not complete optimization. Another critical parameter of chromatographic selectivity, the mobile phase composition, was kept constant during the research. The fact that it is not part of the experimental design limits the possibility of investigating the synergistic effect of pH and temperature. Further research should include a mobile phase composition based on a mixture or mixture-process design capable of multi-factor optimization. Although the quadratic models were statistically significant, there was an overparameterization regarding the information content available. Less hierarchical models were adopted to enhance interpretability, and it is justified to continue testing more extensively under more general conditions.

Author contributions

Asma S. Al‐Wasidi: Conceptualization, Methodology, Writing – Review & Editing. Jumana A. Sanari: Data Curation, Validation, Visualization. Fahad M. Alminderej: Supervision, Formal Analysis, Funding Acquisition, . Sayed M. Saleh: Investigation, Resources, Writing – Final Draft. Hoda A. Ahmed: Software, Writing – Review & Editing. Mahmoud A. Mohamed: Method Development, Experimental Work, Project Administration, Writing – Original Draft.

Funding

The authors declare that no funding was received for this work.

Data availability

Data is provided within the manuscript or supplementary information files.

Declarations

Competing interests

The authors declare no conflict of interest.