Sweet sorghum [Sorghum bicolour (L.) Moench], is known as “heart of the desert” and “camel crop”. Sorghum, the 5^th most significant cereal crop in the world, produces multiple products depending on its variety, such as grain sorghum, which is used for human food. Sorghum stalks are used to produce syrup, alcohol and bioethanol, which might be due to the high amount of sugar accumulation in the stem (Sasy and Abu-Ellail, 2021).
       
Sorghum is the 3^rd largest producer in the world, producing 5.27 million tonnes and shares 8% of the world production after the United States and Nigeria. Leading producing state in India, including Maharashtra (37%), Karnataka (22%), Tamil Nadu (10%), Rajasthan (8%) and Madhya Pradesh (7%) (According to  International Production Assessment Division “IPDA”, 2024). Sweet sorghum contains (Sucrose, fructose, glucose) in stalk and juice percentage share 50-75% of total biomass production and in seed 9-11% protein, 1.2-1.5% minerals, 1.6 fat and 70-72% carbohydrate (Rachman et al., 2025). Sweet sorghum juice has emerged as a vital resource in recent decades, offering significant benefits for human health and environmental sustainability (Amaravel et al., 2023). Juice is a nutritious beverage that provides a natural energy source, supports digestive health and boosts immunity due to its antioxidant properties. Its low glycemic index makes it a healthier alternative to refined sugars, appealing to health-conscious consumers (Olugbemi et al., 2018). Sweet sorghum serves as a dual-purpose crop for food and bioethanol production, supporting renewable energy, reducing greenhouse gas emissions and enhancing food security (Patel et al., 2025). Its juice is a sustainable and health-friendly option, with total sugar content ranging from 12-20% as measured by a Brix refractometer (Reddy et al., 2023). Zeolite improves nutrient retention through high cation-exchange capacity, press-mud enriches soil structure and microbial activity and bio-neem acts as both bio-pesticide and bio-fertilizer to enhance nutrient use efficiency. Integrated use of these inputs strengthens soil health and optimizes nutrient supply, thereby improving sorghum juice quality. FTIR provides a molecular fingerprint of sorghum juice, revealing functional groups linked to sugars, organic acids and phenolics (Dinato et al., 2024; Sharma and Thakur, 2025). Spectral variations reflect nutrient amendments and correlations show integrated nutrient management enhances physical and biochemical juice quality. Thus, FTIR is a rapid, non-destructive tool for monitoring fertilizer-induced quality changes. Taking these things into account, the present study had the following specific goals (i) to evaluate FTIR spectral changes in stalk juice under integrated dose of zeolite, press-mud, bio-neema and nano NPK (ii) to examined the functional group linked to improvements in sugar and bio-chemical quality (iii) to evaluate integrated inputs for their efficiency in enhancing juice composition and overall crop quality.  

Experimental location and design
 
A sweet sorghum study was conducted at the Lovely Professional University research farm, Phagwara, Punjab, India, during the kharif in 2023-24. The experimental site is located at 31°22′31.81 North and a longitude of 75°23′03.02′′ East, with an average elevation of 2 above sea level. Samples were taken at a depth of 0-30 cm as a composite form, then air dried, crushed and proceeded for the analysis. The feature of soil is sandy clay loam texture class including properties: low nitrogen (@ 239.30 kg ha^-1), medium phosphorus (@ 29.3 kg ha^-1), available potassium (@ 247.70 kg ha^-1), OC (0.42%) and pH (7.35). This is a randomized block design; ten different treatment combinations with three-time replicas were considered while applying zeolite, press-mud and bio-neema with or without inorganic fertilizers. Besides, one control treatment with no fertilizers was used. The treatment details are as follows: T1, absolute control; T2, 100% RDF;  T3, 75% RDF + Zeolite (1.5 t ha^-1 ); T4, 75% RDF + Press-mud (15 t ha^-1 ); T5, 75% RDF + zeolite (1.5 t ha^-1) + Press-mud (15 t ha^-1); T6, 75% RDF + zeolite (1.5 t ha^-1) + Press-mud (15 t ha^-1) + Bio-neema (5 kg ha^-1); T7, Nano NPK (19:19:19) + zeolite (1.5 t ha^-1); T8, Nano NPK (19:19:19) + Press-mud (15 t ha^-1); T9, Nano NPK (19:19:19) + Press-mud (15 t ha^-1) + Zeolite (1.5 t ha^-1); T10, Nano NPK (19:19:19) + zeolite (1.5 t ha^-1) + Press-mud (15 t ha^-1) + bio-neema (5 kg ha^-1). The 100% RDF (90:40:40 N, P₂O₅ and K₂O) applied using urea, diammonium phosphate (DAP) and muriate of potash (MOP). Therefore, fertilizers doses were adjusted according to treatment specifications.
 
Juice extraction (%)
 
The juice extraction percentage was calculated as an average by taking an average of five selected plants. Use this equation to determine the ultimate juice %.

 
FTIR analysis on sorghum juice was performed using an FTIR measurement apparatus consisting of a fourier transform infrared spectrophotometer
 
FTIR spectra were measured with a (Model-Spectrum 2 PerkinElmer) outfitted with an ATR attachment. Spectra were acquired in the mid-infrared region of 4000-400 cm^-1 at a spectral resolution of 4 cm^-1 from 32 scans per sample. Background spectra were acquired before each producing run and automatically subtracted to compensate for atmospheric CO₂ and H₂O interference. The collected spectra were processed using (software name, e.g., OMNIC or OriginPro) and then baseline-corrected, normalized and smoothed as necessary. Absorbance bands for critical functional groups, e.g., O-H stretching (3300 cm^-1), C-H stretching (2920 cm^-1), C=O stretching (1740 cm^-1) and C-O-C vibrations (1030 cm^-1) were determined for the sorghum juice, which in turn describes the biochemical behaviour of sorghum juice material. Evidence for the presence of sugars, amino acids, organic acids and phenolic compounds was drawn from standard reference spectra.

Changes in the structure and composition of sorghum juice with T1-to-T4 treatments. Changes in the structure and composition of sorghum juice across T1 to T4 treatments, as observed from FTIR spectra, are presented in Table 1. In general, FTIR peaks are described in wavenumbers (cm^-1) with characteristic stretches such as broad O-H (alcohols, water) at ~3200-3600 cm^-1, C=O (carbonyls) at ~1650-1800 cm^-1, C-H (alkanes) at ~2800–3000 cm^-1, C-O at ~1000-1300 cm^-1 and other groups i.e. N-H, C-N, S=O in the intermediate range between ~1500-1700 cm^-1. Peak signals detected in the spectrum of T1 include 1069.57 and 1406.95cm^-1 (C-O stretch of primary alcohols), 2201.77cm^-1 (N=C=S stretch of isothiocyanate) and 2345.88cm^-1 (O=C=O, possibly CO₂ product) and substantiated the presence of primary alcohol originating from sugars, bioactive and/or degradative isothiocyanates and early fermentation products of CO₂. Spectrum T2 exhibits peaks at 1110.61 cm^-1 (C-O stretching of the secondary alcohol), 1285.32 cm^-1 (COO/CO stretching of ester or acid) and 1657.48 cm^-1 (C=C stretching of alkenes), indicating enzymatic conversion or gentle treatment of the product for phenolic/ester production. Spectrum T3 shows 1314.38 and 1942.20 cm^-1 (C=C=C stretching, allied to allenes), 1541.83 cm^-1 (O-H bending) and 812.15 cm^-1 (C=C bending, alkenes), which imply an oxidative degradation of polyunsaturated compounds, as a result of heat or UV and structural rearrangement, but with the persistence of alcohol sign. 1746 cm^-1 (N=O stretching, nitro compound), 2199.72 cm^-1 (CºC stretching, alkynes) and 2788.88 cm^-1 (O-H stretch) of spectrum T4 indicate that it may be due to advanced microbial fermentation. Also, exposure to a nitrogenous environment led to the production of unusual groups such as nitro and alkyne moieties, possibly arising from the protein or amino acid metabolism. The coatings of spectra shown in combination are indicative of the successive transformation from fresh juice high in sugar (T1), with little further changes, to enzymatically or acid hydrolysed (T2), thermally or oxidatively stressed (T3), to nitrogen enriched or highly fermented states (T4), reflecting the increasingly degree of chemical modification over a range of treatments and storage (Fig 1).




       
The specific spectral peaks and transitions in the FTIR spectra of sorghum juice samples T1, T5, T6 and T7 elucidate and demonstrate a clear view of the chemical transformations occurring under different process conditions. These peaks are attributed to other functional groups and represent specific reactions, such as degradation, crosslinking and conformational changes of biomolecules in the juice matrix, characteristic of various beverages. In the FTIR spectrum of T1, a strong peak at 1069.57 cm^-1 is present due to the C-O stretching vibration (primary alcohol), which suggests that unmodified carbohydrates such as glucose and fructose existed (Zeinalipour-Yazdi and Loizidou, 2021). The peak at 1406.95 cm^-1 is evidence for the existence of polysaccharide structures. An individual peak at 2201.77 cm^-1, attributed to N=C=S stretching vibrations, might represent an early decomposition product of sulfur-containing compounds (e.g., glucosinolates or thiol precursors).
       
A band indicated the initial microbial fermentation at 2345.88 cm^-1, a signature of CO₂ vibration. These attributes demonstrate that T1 retains primarily chemical integrity controlled by native sugars and early biochemistry and with limited structural degradation or crosslinking. In T5, C=O and C-O-C stretching bands at 1104.46 and 1140.62 cm^-1 suggest converting sugars to esters and aldehydes (mild oxidations or even Maillard-type reactions). A peak at 2479.82 cm^-1 is characteristic of S-H stretching and may indicate sulfur-based degradation or thiol formation from enzymic or oxidative cleavage. The spectrum evidences the formation of new chemical links (e.g., esters) and a minor degradation, even if the molecular integrity is exceptionally well preserved. Sample T6 presents a strong C=C stretching band at 1632.35 cm^-1, representing the conjugated alkenes commonly produced during unsaturated compounds’ thermal degradation or photochemical rearrangements (Amanat et al., 2021). A wide O-H stretching band at 3298.56 cm^-1 which is associated with the O–H stretching band at 3737.99 cm^-1 of the secondary OH group, is evidence of the existence of intermolecular bonded polymeric alcohols and possible hydrogen bonding in the thermally modified polysaccharides. Both these peaks indicate a significant degree of structural rearrangement, cross-linking and retention of alcohols even at high-energy irradiation. The occurrence of a benzene derivative band at 622.77 cm^-1 and a strong C-O and C=C stretching band at 1368.97 cm^-1, which is detected in T7, reveals the formation of an aromatic compound that might have resulted from the advanced microbes’ degradation or the synthesis of secondary metabolites. The presence of a C=O stretching band at 2348.49 cm^-1 indicates the transformation of primary metabolites into more oxidized or aromatic forms. FTIR profiles across treatments reveal a clear progression from T1, dominated by simple carbohydrates, to T7 and beyond, characterized by more complex and oxidized structures. Early pyrolysis stages largely preserve structural integrity with limited decomposition, whereas advanced processing promotes the formation of aromatic rings, conjugated alkenes and oxidized functional groups. Evidence of ester formation and polyhydroxy stretching, particularly in T6, further suggests cross-linking and structural reorganization, contributing to enhanced bioactivity and sensory attributes. The 2271, 2534 and 3291 cm^-¹ peaks represent CO₂, S-H, alkynes/alkenes of degradation, showing enzymatic or thermal breaking of the native bonds. These structural changes indicate that processing might modulate the sorghum juice molecular fingerprint, thus enhancing or tailoring its nutritional, functional and sensory properties according to the intended application (Fig 2).


       
Structural and functional group changes in the samples before and after the treatment were examined using Fourier Transform Infrared (FTIR) spectroscopy. The FTIR spectra of untreated (T1) and treated (T8, T9 and T10) samples. For the untreated control (T1) the FTIR spectrum showed the main absorptions at 1046 cm^-1 (C-O stretch of primary alcohols), 2921 cm^-1 (C-H stretch of alkanes), 2345 cm^-1 (O=C=O stretch) for which the spectra was not observed for the extracted samples, as well as 2141 cm^-1 (N=C=S stretch) typical of an unmodified organic matrix, most likely natural polymers such as cellulose or lignin derivatives commonly found in plants or biomass derived materials. After treatment, a significant increment in the bands of sample T8 was observed, with shifts in alkane bands to 2914 cm^-1 and 1546 cm^-1 attributed to N=O stretching (nitro compound). This moiety’s persistence and/or chemical stability in such matrices is observed as the N=C=S function persists at 2141cm^-1. The presence of nitro functionalities hints that oxidation or nitrative events have occurred during treatment, which may enhance the material’s reactivity or potential antimicrobial activity. A sample, such as T9, exhibited either new or shifted functional groups, such as a band at 1141 cm^-1 (C-O stretching of secondary alcohols), in contrast to the primary alcohol in T1. Moreover, a new absorption at 2571 cm^-1 assigned to S-CN (thiocyanate) indicated the introduction of sulfurated groups (Novovic et al., 2019). The observed spectral changes suggest possible sulfonation or thiolation, imparting enhanced bioactive or degradative properties to the processed sample. Treatment T10 showed a distinct chemical profile, with new FTIR bands at 874 cm^-1 (alkene C-C bending), 1258 cm^-1 (C-O stretching of esters/ether linkages) and a broad band near 2503 cm^-1 attributed to O-H stretching of carboxylic acids. These features indicate depolymerization and/or oxidative cleavage of macromolecules, leading to the formation of unsaturated and acidic functional groups that may enhance solubility, biodegradability and binding potential. The untreated samples mainly comprised hydroxyl, alkyl and isothiocyanate groups. In contrast, the treated samples presented a more heterogeneous profile of functional groups, such as nitro (T8), thiocyanate (T9) and alkene/carboxylic acid functionalities (T10). Change from primary to secondary alcohols and the generation of reactive functional groups show a major conversion of the base matrix material. Such changes indicate the better physicochemical properties of the treated biomaterials, which are useful for certain targeted applications (Fig 3).


       
Fig 4 is a PCA biplot,  showing the multivariate grouping of FTIR spectral data as a function of different sample treatments by dimension reduction from PC1 and PC2 (99.94% of the total variance was explained by PC1, which means it explains almost the entire variability in the data). The data is grouped by confidence interval/variation of the respective spectral fingerprints and is depicted as elliptical data points on the plot with different colours for each group. Especially for some groups, the data points were closely packed near the origin, indicating that these samples had similar chemical composition or exhibited slight variation in FTIR spectra. On the other hand, the more elongated ellipses (e.g. red and blue ones) reveal higher variance or other different chemical changes introduced by the treatments, probably related to the significant structural changes observed in FTIR analysis (e.g., nitro, thiocyanate, carboxylic) (Cao et al., 2012). The horizontal spread of PC1 along the wavelength axis indicates that major spectral variation arises from differences in chemical bond vibrations across the wavenumber range. Overall, the PCA confirms chemically distinct treatment groupings, consistent with FTIR peak shifts showing that modifications such as oxidation, sulfonation and esterification significantly altered the molecular structure of the native matrix.


       
PCA analysis shows PC1 (Principal component 1) explains 99.94% of the variance of the original data, with the first PCA component data, with an eigenvalue equal to 9.994285, hence the most non-trivial variation observed in the data. PC2 (Principal component 2) accounts for 0.05% of the variance (eigenvalue 0.004933), indicating it accounts for very little additional explanatory power. Each treatment’s loading coefficients (T1 to T10) indicate the measurements’ contributions to PC1 and PC2. T1 scores loading of 0.316115 (PC1) and 0.504739 (PC2) strongly influenced both components. T2 contributes 0.316303 (PC1) and -0.09292 (PC2), indicating a weak negative effect on PC2. T3 impacts 0.316151 (PC1) and -0.44253 (PC2), with a weak negative effect on PC2. T4 has values of 0.316302au (PC1) and 0.001072au (PC2), which have almost no relationship with PC2. T5 is distinguished by 0.316056 (PC1) and 0.574733 (PC2), which falls far out in PC2. For T6, it is 0.316196 (PC1) and -0.38117 (PC2), indicating a considerable negative effect on PC2. T7 provides 0.316298 (PC1) and -0.08703 (PC2), a slight negative effect in PC2. T8 contributes 0.316298 (PC1) and -0.01178 (PC2), nearly neutral in PC2. T9: 0.316290 (PC1) and 0.134004 (PC2), with a small positive impact on PC2. T10 contributes 0.316268 (PC1) and -0.19873 (PC2), indicating moderate negative loading on PC2. In general, all the treatments exhibit similar loadings on PC1 (around 0.316), indicating no significant difference as to their importance in explaining the main variance in the data set. The difference between treatments is more apparent in their PC2 loadings, which are not only of different signs but also appear at a second level of differentiation (Table 2).


       
Fig 5 is a 4-panel plot from PCA that has proven to be illustrative of the multivariate structure in a dataset. The scree plot (top left) shows a substantial decrease after principal component 1 (PC1), for which only the corresponding eigenvalue is larger than 10 and the others have approximately zero eigenvalues (Bagewadi et al., 2016). PC1, the dominant axis, accounts for nearly all variability in the dataset. The loading plot shows that most treatments (T1-T10) contribute almost equally to PC1 (~0.316), indicating a strong common trend. Although PC2 does not clearly separate treatments, slight positive loadings for T5 and T1 and negative loadings for T3 and T6 suggest subtle, treatment-specific variations. The scores plot (bottom, left), which shows the projection of individual observations onto PC1 and PC2 axes, also reinforces the interpretation that PC1 captures the most variation (horizontal spread and vertical dispersion).


       
The combined biplot (bottom right) is adapted to this, which maps treatment loadings onto the scores of single observations and emphasises that almost all structural differentiation occurs along PC1. These visual trends are confirmed by the quantified explained variance in the analysis: PC1 explains 99.94% of the total variance and PC2 only 0.05%. The relatively uniform contribution of treatments to PC1 indicates a strong common trend across the data, while minor separation along PC2 particularly for T5 and T3 suggests subtle treatment-specific effects. Overall, the PCA reveals a predominantly unidimensional structure with a dominant primary component and limited but meaningful variation in the second component (Cheng 2022). This reflects a strong and consistent treatment response, while indicating that treatments with distinct PC2 loadings may offer biologically or experimentally meaningful insights.
       
There was a wide range in juice yield across treatments, which may be associated with differences in biological and physiological mechanisms related to nutrient management, soil amendments and plant bio-stimulation. The latter is in agreement with the control treatment, the one with the smallest yield (9797.58 l/ha) and that shows the low physiological activity when plants grow under a nutrient-stressed situation, with photosynthetic limitations, the root system development and the partition of the assimilates. Juice yield was significantly increased by 100% RDF (30246.88 l/ha) condition, suggesting that the balanced macro-nutrient availability (90:40:40 NPK) is essential for chlorophyll synthesis, enzymatic activity and energy metabolism. These contribute to vigorous vegetative growth, higher LAI, enhanced net photosynthetic rate, better translocation of the assimilates into the sink organ (stem) and the accommodation of reserve materials (juice-storing tissues).
       
Additionally, application of 75% RDF with zeolite or press-mud, isolated, retained yields (28245.14 and 29102.02 l/ha, respectively) close to FRDF, indicating the effectiveness of these amendments in enhancing plant physiological processes. Zeolite, due to its high cation exchange capacity, enables slow and sustained nutrient release, reducing leaching and maintaining prolonged stomatal conductance and photosynthetic activity. In contrast, press-mud enriches organic matter and microbial diversity, enhancing root vigor, nutrient and water uptake, cell turgor and ultimately juice accumulation. The maximum juice yield was recorded in 75% RDF + zeolite + press-mud + Bio-neema (36100.76 l/ha), indicating synergistic effects on physiological and metabolic activity. Bio-neema- Containing chemical compounds such as azadirachtin and other limonoids- was presumably a bio-stimulant and mild elicitor, possibly modulating phytohormones such as auxins and cytokinins. This may enhance cell division, delay senescence and improve assimilate partitioning to stem tissues. The combined use of press-mud and zeolite improved root–soil contact and root longevity, enhancing water and nutrient uptake, leaf water status, stomatal function and photosynthate accumulation, which in turn increased juice extraction through improved sucrose synthesis, loading and storag (Jetti et al., 2024). This is supported by an improved source sink relationship and enhanced activity of sugar-metabolizing enzymes, particularly sucrose phosphate synthase and invertase. In contrast, Nano NPK (19:19:19) treatments produced lower yields (17034.40 to 20849.46 l/ha), suggesting that while nano-fertilizers ensure rapid nutrient availability, their effects are short-lived without complementary amendments. Limited sustained mineral release and slower soil biological and structural improvements may restrict long-term physiological efficiency, especially during juice filling and ripening stages (Shinde et al., 2025). Overall, the results highlight the potential for partial substitution of chemical fertilizers with zeolite, press-mud and bio-neema, which not only sustains but enhances plant physiological functions, nutrient-use efficiency, hormonal balance, root development, photosynthate accumulation and ultimately improves juice yield and quality (Table 2).

The FTIR analysis of sorghum juice provided a rapid, non-destructive and informative fingerprint of its biochemical composition. The spectra revealed characteristic absorption bands corresponding to functional groups associated with carbohydrates (notably sucrose and glucose), organic acids, amino compounds and phenolic constituents. The dominant peaks in the regions of 3200-3600 cm^-1 (O-H stretching), 2800-3000 cm^-1 (C-H stretching) and 1000-1200 cm^-1 (C-O stretching) confirmed the presence of major hydrophilic and bioactive constituents typically found in sorghum stalk juice. The high resolution and sensitivity of FTIR allowed for the qualitative differentiation of key molecular signatures without the need for complex sample preparation. Overall, FTIR proved to be an efficient and valuable analytical technique for characterising the molecular composition of sorghum juice, supporting its use in food, nutraceutical and bioethanol production applications. These insights also facilitate future comparative studies across genotypes, treatments, or processing methods in sorghum research.

The authors thank their supervisor for guiding them and the School of Agriculture, including the central instrumental library of Lovely Professional University, for providing the platform to conduct this research.
 
Disclaimers
 
The views and conclusions expressed in this article are solely those of the authors and do not necessarily represent the views of their affiliated institutions. The authors are responsible for the accuracy and completeness of the information provided, but do not accept any liability for any direct or indirect losses resulting from the use of this content.

The authors declare there is no competing interest.