Potential Application of Nanocellulose Derived from Bagasse and Durian Rind for Mitigation of Mycotoxin Contamination in Poultry Diets

Abstract

This study was aimed at producing nanocellulose from sugarcane bagasse and durian rind residues for applications to determine adsorption capacity against mycotoxin in poultry diets. Durian rind-derived nanocellulose exhibited finer fiber (12–21 nm diameter and 197–350 nm length) and higher yield (42.1%) than bagasse-derived nanocellulose (18–36 nm diameter and 82–169 nm length), with FTIR confirming purer cellulose I/II structures. The in vitro test adsorption capacity against ochratoxin (OTA) was determined at an incubation time of 180 min to establish working conditions. It was found that the working conditions of bagasse-derived nanocellulose and durian rind-derived nanocellulose were 33 mg/mL and 36.5 mg/mL, respectively. Subsequently, using these working conditions, adsorption capacity was determined via an in vitro digestibility test. Bagasse-derived nanocellulose exhibited an adsorption capacity against OTA of 35.59%, while durian rind-derived nanocellulose achieved an OTA adsorption rate of 39.53% at a contact time of 3 h. Naturally contaminated poultry feeds collected from nine farms in Chiang Mai, Thailand, indicated that both types of nanocelluloses achieved minimum–maximum OTA adsorption rates of 42–43%, aflatoxin B1 (AFB1) at 29–30%, and fumonisin B1 (FB1) at 21–23% across the nine farms’ mean values. These findings suggest that nanocellulose derived from sugarcane bagasse and durian rind has potential as a sustainable biosorbent for improving mycotoxin management in poultry production.

1. Introduction

Agricultural residues or byproducts derived from crop harvesting and processing, such as bagasse and durian rind, are currently generated at a rate of millions of tons a year in Thailand. However, the limitation of materials possessing utility value to farmers are consequently disposed of through field incorporation or combustion. Remarkably, the burning of agricultural residues constitutes a major contributor to particulate matter (PM 2.5) pollution in Northern Thailand, with significant implications for public health. Therefore, the utilization of agricultural residues represents a critical solution to mitigate burning practices while simultaneously adding value to agricultural products. Furthermore, these agricultural residue materials mainly comprise lignocellulose, which is a principal constituent of plant cell walls consisting of hemicellulose and cellulose and comprises approximately 40–70% of plant cell walls [1]. Cellulose is a crucial component of all plant materials, providing structural integrity to plant cell walls. It consists of linear chains of homopolysaccharides made up of β-D-glucopyranose units interconnected by β-1-4-glycosidic linkages. The primary constituents are carbon, hydrogen, and oxygen [2,3,4]. In the subunit of cellulose, nanocellulose is interestingly remarkable due to the properties of nanocellulose, such as its small dimension, high surface area, variety of shapes, and large capacity to hold water [4]. Recent advances in science and technology have enabled the extraction and size reduction of cellulose to produce nanocellulose, which represents an alternative material derived from natural sources. Nanocellulose exhibits exceptional mechanical properties, a high specific surface area, and elevated reactivity, thereby enabling the reinforcement of composite materials and demonstrating a pronounced capability for water adsorption and retention. Notably, nanocellulose is biodegradable and exhibits natural decomposition characteristics [5].

Currently, the livestock production increase simultaneously responded to rises in population. One of those kinds is poultry production, which is the majority of all livestock production. However, the biggest problem in poultry production is mycotoxin contamination in diet. Generally, poultry diets are very frequently contaminated with two or more mycotoxins [6]. A recent report indicated that 88% of the tested corn samples from the United States were positive for at least one mycotoxin, and over 90% of the tested poultry feed ingredients had more than one mycotoxin [7]. The mycotoxins most commonly found in foods and feeds have been determined to be aflatoxins (AF: B1, B2, G1, and G2); zearalenone (ZEA); deoxynivalenol (DON); T-2 toxin (T-2); fumonisins (FUM: FB1, FB2, and FB3); patulin; and ochratoxin A (OTA) [8]. Mycotoxin intake causes health problems in poultry, such as decreased feed intake and production drops, as well as the interruption of normal biological processes and organs, including those of the liver, kidney, and intestine [7,8]. In the past decade, several studies have tried to mitigate the harmful consequences of mycotoxin contamination. Many strategies have been explored to degrade, eliminate, inactivate, and extract mycotoxins contaminating feed, including chemical, physical, and biological methods. However, the many previous methods used often have side effects and drawbacks for food quality and safety. Nanocellulose may potentially solve this point because it has been proven to be less harmful to cells, with a cell viability rate of 80–100% [4]. Previous research presented that nanocellulose from bagasse conjugated with polylysine has the potential to absorb 94.4–100% fumonisin B1 at a concentration of 2000–200,000 µg/mL [4]. However, the nanocellulose from agricultural residue for the adsorption of mycotoxin in poultry diets has yet to be investigated. Therefore, the aim of this study was to extract nanocellulose from agricultural waste such as sugarcane bagasse and durian rind to evaluate the potential for the adsorption of mycotoxin contamination in poultry diets.

2. Results and Discussion

2.1. Characterizations of Nanocellulose

Cellulose extracted from bagasse and durian rind exhibited similar lightness (L* ≈ 73) and whiteness indices (WI ≈ 72). However, following acid hydrolysis, the lightness and WI of the nanocellulose samples decreased slightly, reflecting mild darkening typically associated with the partial degradation of residual non-cellulosic components [9]. Durian rind-derived nanocellulose demonstrated a lower yellowness index (YI = 10.2) relative to that of bagasse (YI = 11.5), suggesting more effective removal of non-cellulosic components during the bleaching process [10].

The extraction yield of cellulose differed significantly between raw materials, with bagasse yielding 52.3% compared to 33.1% from durian rind. Interestingly, durian rind produced more nanocellulose (42.1%) than bagasse (32.4%), probably because its amorphous regions break down more easily during acid hydrolysis. Furthermore, the ΔE values increased significantly after acid hydrolysis, reflecting the perceptible color shift associated with structural modification and the removal of colored non-cellulosic residues [4]. Durian rind nanocellulose showed a higher ΔE (6.2), indicating a more apparent color change compared with bagasse (4.4).

Figure 1 shows the FE-SEM images of nanocellulose obtained from durian rind and bagasse. Durian rind-derived nanocellulose exhibits a smaller and more elongated rod-like structure compared to that of bagasse. The durian rind nanocellulose presents finer diameters ranging from approximately 12–21 nm and substantially longer fibrils of about 197–350 nm (Figure 1a). In contrast, nanocellulose derived from bagasse shows larger diameters of roughly 18–36 nm and shorter fibrils ranging from 82 to 169 nm (Figure 1b). These observations highlight the strong relationship between the aspect ratio, morphological characteristics, and the resulting performance of nanocellulose derived from durian rind. Consistent with earlier reports, durian rind-derived nanocellulose exhibits superior properties compared to those of bagasse-derived nanocellulose, reflected in its cellulose I structure, higher aspect ratio, and enhanced viscosity and swelling behavior [4,9]. Bagasse-derived nanocellulose exhibited shorter rods and a greater tendency toward aggregation, suggesting that residual hemicellulose and lignin may have inhibited complete fibrillation during acid hydrolysis, as previously observed in other lignocellulosic materials [11]. These findings agree with earlier reports showing that variability in nanocellulose morphology (length and diameter) strongly depends on the chemical composition and microfibrillar arrangement of the parent biomass [12]. In particular, biomass with higher lignin content or more complex cell–wall architecture often requires more severe hydrolysis to achieve fibrillation, which may otherwise produce shorter nanocellulose material or cause aggregation if insufficiently treated [13,14].

2.2. Chemical Characterization of Nanocellulose

The chemical structure of durian rind-derived nanocellulose and bagasse-derived nanocellulose was characterized by FTIR. Figure 2 shows the FTIR spectra of nanocellulose extracted from durian rind and bagasse. The broad absorption band at 3340–3440 cm⁻¹, corresponding to O–H stretching vibrations, was observed in both nanocellulose samples. Specifically, the peaks at 3490 and 3444 cm⁻¹ represent hydrogen-bonded O–H groups characteristic of cellulose I and II [15]. These peaks indicate the presence of both intra- and intermolecular hydrogen bonding [16,17]. This is consistent with previous studies, which reported that durian rind-derived nanocellulose exhibits a higher content of cellulose I-type intermolecular hydrogen bonds compared with that of bagasse-derived nanocellulose. In addition, nanocellulose shows a slightly stronger peak at 3490 and 3444 cm⁻¹ of intermolecular hydrogen bonds related to cellulose II [4]. Peaks around 2900 cm⁻¹ associated with C–H stretching of –CH and –CH₂ groups further support the presence of cellulose structures and the absence of aromatic C–H vibrations from lignin [2]. The 1635 cm⁻¹ band corresponds to the O–H bending vibration of the absorbed water of nanocellulose. Additionally, nanocellulose exhibits more intense peaks at 1420, 1376, 1316, 1160, 1060, and 895 cm⁻¹, which are associated with CH₂ bending, C–H bending, C–O–C stretching, and β-1,4-glycosidic linkages. The stronger intensities of these cellulose-specific bands indicate that durian rind nanocellulose possesses higher preservation of the cellulose backbone and a more ordered molecular structure than bagasse nanocellulose [18]. These FTIR results are consistent with the yield data (

Table 1. Color parameters (L*, a*, b*, WI, ΔE, and YI) and extraction yield (%) of cellulose and nanocellulose obtained from durian rind and bagasse.

Type of Nanocellulose	L*	a*	b*	WI	ΔE	YI	Yield (%)
Cellulose bagasse	72.8 ± 0.1 a	1.6 ± 0.0 a	7.0 ± 0.1 a	71.9	0.0	13.7	52.3 ± 1.5 a
Cellulose durian rind	73.0 ± 0.2 a	1.5 ± 0.0 b	7.3 ± 0.0 b	72.0	0.0	14.2	33.1 ± 1.2 b
Nanocellulose bagasse	68.4 ± 0.5 a	1.8 ± 0.1 a	7.5± 0.1 a	67.5	4.4	15.8	32.4 ± 0.2 a
Nanocellulose durian rind	67.3 ± 0.2 a	1.5 ± 0.0 b	4.8 ± 0.2 b	66.9	6.2	10.2	42.1 ± 0.6 b

* Values are mean ± SD. Different letters (^a and ^b) within a column and within the same group indicate significant differences (p < 0.05).

), where the slightly lower yield of durian rind nanocellulose suggests a more extensive removal of non-cellulosic components, leading to higher cellulose purity compared with bagasse.

2.3. In Vitro Mycotoxin Adsorption of Nanocellulose

The potential adsorption capacities of the nanocelluloses were determined by first testing for the optimal conditions. In Figure 3 is shown the bagasse-derived nanocellulose (Figure 3A) at different concentrations, which exhibited similar mycotoxin adsorption capacities when incubated for 30 and 60 min, with a marked difference observed after 180 min of incubation. Among the tested concentrations, nanocellulose at 33 mg/mL demonstrated the highest ability to adsorb mycotoxins. Meanwhile, nanocellulose derived from durian rind (Figure 3B) exhibited a progressively increased capacity to adsorb mycotoxins as both the concentration of nanocellulose and the incubation period increased. Durian rind-derived nanocellulose exhibited finer fibril diameters and a higher aspect ratio than bagasse-derived nanocellulose, which likely translates to a larger specific surface area and more interconnected fibrillar network, resulting in more accessible blinding sites and a progressive increase in mycotoxin adsorption with increasing concentration and incubation time [19]. Conversely, bagasse-derived nanocellulose forms shorter rods with a stronger tendency to aggregate. This probably reduces the effective surface area and the number of available adsorption sites, leading to an earlier adsorption plateau and small gain in mycotoxin removal at prolonged incubation times. This is in agreement with a previous study showing that differences in the fibrillation degree and aggregation behavior among lignocellulosic sources strongly influence adsorption affinity and the kinetics of nanocellulose adsorbents [4].

Based on preliminary experiments, the highest workable nanocellulose concentration was selected as the working condition and an incubation time of 180 min was chosen as a sufficient contact time for nanocellulose–mycotoxin interaction. This decision was made to ensure optimal conditions for evaluating the adsorption capacity of nanocellulose toward mycotoxins. The extended incubation period allows for sufficient time for adsorption equilibrium to be reached, thereby providing reliable and reproducible data for subsequent analyses. The adsorption of mycotoxins by nanocellulose presented only minor differences between 30 and 60 min, whereas a more pronounced increase was observed at 180 min. The result suggests a diffusion-controlled process in which toxins require sufficient time to mitigate into the internal pores and fibrillar network before equilibrium is reached [20,21,22]. This behavior is consistent with kinetic studies of bio-based adsorbents for OTA and AFB1, where adsorption commonly follows pseudo-second-order kinetics and is governed by intra-particle diffusion, leading to a relatively slow approach to equilibrium at extended contact times. The choice of 180 min as the standard contact time for subsequent experiments in buffer and in the in vitro digestion model is justified, as it ensures that adsorption is evaluated under near-equilibrium conditions, providing more reliable comparisons of nanocellulose performance across treatments [21,23].

To determine the optimal incubation period for the adsorption capacity of nanocellulose, we evaluated the performance of both bagasse- and durian rind-derived nanocellulose using the in vitro digestion model that simulates the avian gastrointestinal tract. During these experiments, various incubation times were tested to assess their effect on the adsorption efficiency of mycotoxins by nanocellulose present in contaminated feed samples (Figure 4). The result showed that various incubation times did not lead to a difference in the adsorption of ochratoxin in the in vitro digestion model (p > 0.05). Therefore, the duration of incubation time was 3 h. To evaluate the effectiveness of nanocellulose in adsorbing ochratoxin from poultry feed under simulated digestive conditions, both bagasse-derived and durian rind-derived nanocelluloses were tested using the established in vitro digestion model.

In the in vitro digestion model simulating the poultry gastrointestinal tract, there was no significant time-dependent differences in OTA adsorption. This can be attributed to the more complex physicochemical environment (pH, ionic strength, and the presence of bile salts and dietary proteins) compared with simple buffer systems, a condition known to accelerate the attainment of quasi-equilibrium and to partially block adsorbent surfaces in mycotoxin binding studies [24,25]. Under this condition, a 3 h incubation time is considered sufficient to encompass the effective residence time of digesta in the proventriculus, gizzard, and small intestine of broilers and has therefore been widely adopted as a standard contact time in in vitro gastrointestinal models for evaluating mycotoxin binders in poultry diets [26,27].

The results demonstrated that nanocellulose sourced from durian rind exhibited a higher adsorption efficiency for ochratoxin when compared to bagasse-derived nanocellulose. Specifically, the OTA adsorption rate achieved with durian rind nanocellulose was 39.53%, whereas bagasse-derived nanocellulose achieved an adsorption rate of 35.59%; however, the difference was not statistically significant (p > 0.05). These findings, as illustrated in Figure 5, highlight the superior performance of durian rind nanocellulose in mitigating ochratoxin contamination in feed samples during in vitro digestion. The result indicated that the structural advantages of durian rind-derived nanocellulose such as a higher aspect ratio, higher cellulose purity, and a more developed hydrogen-bonding network remain effective even in the presence of competing feed matrix components [4,28]. This observation is consistent with the FTIR results showing better preservation and ordering of the cellulose backbone in durain rind-derived nanocellulose. This likely increases the density of accessible hydroxyl groups and specific binding domains capable of forming hydrogen bonds and hydrophobic/π–π interactions with OTA’s carboxyl and aromatic moieties. This mechanism has been proposed for cellulose-based and polysaccharide adsorbents targeting OTA in complex food and feed systems [2,20,28].

2.4. Application of Nanocellulose for Mitigation of Mycotoxin Contamination in Poultry Diet

Following a laboratory evaluation of nanocellulose adsorption efficiency, a real-world application was tested using feed samples collected from nine commercial poultry farms in Chiang Mai, Thailand. These feeds were analyzed for contamination by three major mycotoxins: OTA, AFB1, and FB1. Poultry feed samples from nine commercial farms exhibited mycotoxin contamination level of OTA at 0.33–1.44 mg/kg, AFB1 at 5.54–30.25 µg/kg, and FB1 at 2.33–53.95 mg/kg. Notably, OTA levels frequently exceeded the EU guidance value of 0.5 mg/kg (Directive 2002/32/EC), while AFB1 remained below the strict 20 µg/kg maximum limit. In terms of FB1, levels approached but did not exceed the 50 mg/kg threshold (5 mg/kg recommended) for poultry feeds. The application of bagasse-derived nanocellulose showed different levels of adsorption of the three common mycotoxins found in poultry feed. The adsorption of OTA in the nine farm samples ranged from 29.07% to 61.43% (43.03% ± 9.58%). The adsorption of AFB1 ranged from 15.10% to 40.48% (29.20% ± 7.87%), and the adsorption of FB1 ranged from 11.74% to 28.10% (21.05% ± 6.59%). OTA had the highest average adsorption efficiency, followed by AFB1. FB1 had the lowest overall adsorption (

Table 2. Efficiency of bagasse-derived nanocellulose in adsorption of mycotoxin contamination in poultry diet. OTA, ochratoxin; AFB1, aflatoxin B1; FB1, fumonisin B1.

Farms	OTA (mg/kg Feed)			AFB1 (µg/kg Feed)			FB1 (mg/kg Feed)
	Before	After	%Adsorption	Before	After	%Adsorption	Before	After	%Adsorption
1	0.37	0.24	34.52	27.01	22.93	15.10	53.95	41.97	22.20
2	0.86	0.61	29.07	13.68	11.12	18.72	4.90	3.68	24.71
3	1.20	0.63	47.09	30.17	20.85	30.90	16.90	13.51	20.03
4	1.18	0.70	40.51	21.88	14.98	31.54	10.28	7.47	27.31
5	0.72	0.41	43.60	5.88	3.85	34.49	16.00	11.99	25.04
6	0.77	0.47	39.23	12.55	8.86	29.34	9.99	7.18	28.10
7	1.05	0.60	42.34	10.72	7.33	31.52	2.44	2.16	11.74
8	1.39	0.70	49.44	10.08	6.98	30.75	34.70	30.59	11.83
9	1.06	0.41	61.43	10.08	6.00	40.48	7.71	6.28	18.54
Average	0.95	0.53	43.03	15.79	11.43	29.20	17.43	13.87	21.05

). The significant variation in adsorption efficiency from one farm to another indicates that the baseline levels of mycotoxin contamination, the composition of the feed, or the consistency of the application of nanocellulose may differ from site to site. Similarly, durian rind-derived nanocellulose exhibited a comparable pattern of mycotoxin adsorption. The adsorption of OTA was between 26.93% and 58.88% (43.02% ± 8.51%), the adsorption of AFB1 was between 15.75% and 39.59% (29.84% ± 8.85%), and the adsorption of FB1 was between 15.13% and 32.42% (22.93% ± 4.96%). OTA had the highest average adsorption efficiency, followed by AFB1, and FB1 had the lowest overall adsorption (

Table 3. Efficiency of durian rind-derived nanocellulose in adsorption of mycotoxin contamination in poultry diet. OTA, ochratoxin; AFB1, aflatoxin B1; FB1, fumonisin B1.

Farms	OTA (mg/kg Feed)			AFB1 (µg/kg Feed)			FB1 (mg/kg Feed)
	Before	After	%Adsorption	Before	After	%Adsorption	Before	After	%Adsorption
1	0.37	0.27	26.93	27.01	22.75	15.75	53.95	42.98	20.32
2	0.86	0.49	42.79	13.68	10.53	23.04	4.90	3.97	18.96
3	1.20	0.75	37.26	30.17	18.87	37.46	16.90	12.77	24.38
4	1.18	0.67	42.70	21.88	16.42	24.97	10.28	7.91	23.07
5	0.72	0.38	46.77	5.88	3.53	39.59	16.00	12.52	21.77
6	0.77	0.42	45.37	12.55	7.72	38.40	9.99	7.42	25.65
7	1.05	0.60	42.42	10.72	7.38	31.02	2.44	1.84	24.69
8	1.39	0.57	58.88	10.08	6.84	32.13	34.70	29.45	15.13
9	1.06	0.59	44.07	10.08	6.10	39.46	7.71	5.20	32.42
Average	0.95	0.53	43.02	15.79	11.13	31.31	17.43	13.78	22.93

). The large differences between farms suggest that there are different levels of baseline contamination, consistency in nanocellulose processing, and effects on the composition of the feed matrix.

Consistent across both nanocellulose sources, OTA exhibited the highest mean adsorption (42–43%), followed by AFB1 (29–31%), whereas FB1 showed the lowest overall adsorption (21–23%). This result indicated that the interaction between nanocellulose and each mycotoxin depends strongly on toxin structure and polarity. This pattern agrees with previous reports that polysaccharide- and nanocellulose-based material tend to bind OTA and AFB1 more efficiently than FB1 via a combination of hydrogen bonding, van der Waals forces, and hydrophobic regions, whereas the highly hydrophilic polyanionic nature and extended aliphatic chain of FB1 reduce its affinity for unmodified biopolymer surfaces [25,27]. The close similarity of the mean adsorption between bagasse-derived and durian rind-derived nanocellulose in real feed showed only a modest advantage of durian rind-derived nanocellulose for FB1 and minor differences for OTA and AFB1. This indicated that the initial contamination level plays a dominant role in determining field performance, attenuating the structural differences observed under simplified in vitro conditions. This study is consistent with in vitro–in vivo translation studies showing that matrix components can mask or compete for binding sites on biopolymer adsorbents, thereby reducing the impact of the adsorbent source while emphasizing the importance of the physicochemical environment. From a practical perspective, the average reductions of approximately 30–40% for OTA and AFB1 in poultry diet place these nanocellulose preparations within the performance range reported for other biosorbents and fiber-based binders proposed for mycotoxin mitigation in livestock nutrition [25,27]. Consequently, nanocellulose derived from locally available agriculture wastes such as bagasse and durian rind may serve as promising complementary mycotoxin binders in poultry diet, particularly in regions where synthetic binders are costly or less accessible, aligning with current interest in sustainable bio-based adsorbent material for feed safety.

3. Conclusions

This study demonstrates the potential of nanocellulose derived from agricultural residues such as sugarcane bagasse and durian rind as novel biosorbents for mitigating OTA, AFB1, and FB1 contamination in poultry diets. Both materials exhibited substantial mycotoxin adsorption (30–45%) under in vitro gastrointestinal conditions and in real feed samples from commercial farms. The structural advantages of durian rind-derived nanocellulose confers modest superiority. These findings represent a sustainable approach to feed safety using locally abundant lignocellulosic wastes for mitigating the contamination of mycotoxin in feed. However, the in vivo experiment warrants validation and mechanistic studies in poultry production.

4. Materials and Methods

4.1. Materials

Two types of cellulose from durian rind and bagasse were obtained using the extraction method reported in our previous study [4]. Sodium hydroxide and sulfuric acid, all of analytical reagent (AR) grade, were procured from RCI Labscan Ltd. (Bangkok, Thailand).

4.2. Preparation of Nanocellulose

The method was modified from our previous report [4]. In brief, 50 g of dry cellulose basis of lignin-free cellulose isolated from durian rind and bagasse was hydrolyzed at 50 °C for 5 h with 32% (v/v) sulfuric acid, and the reaction was stopped by adding distilled water. The combination was centrifuged at 4500 rpm for 15 min and rinsed several times, followed by pH adjustment to 7 with 1% (w/v) sodium hydroxide. The neutralized mixture was blended at 36,000 rpm for 15 min and left to stand overnight. It was then blended once more and ultrasonicated in an ice bath (20 kHz, 700 W) for 30 min. The resulting suspension was centrifuged again under the same conditions until no solid residue remained, and an additional sonication was applied to prevent overheating for 30 min. Finally, the nanocelluloses were stored at 4 °C for further use, while a portion was freeze-dried at −50 °C and 20 Pa for 24 h. The resulting freeze-dried powders were used for physical and chemical structural characterization. The nanocellulose yield from both biomass sources was calculated based on the mass of freeze-dried nanocellulose according to Equation (1), following the method reported by Thipchai et al. [4].

$$\mathrm{Y}\mathrm{i}\mathrm{e}\mathrm{l}\mathrm{d} (\%)=W_{nanocellulose}/W_{cellulose} \times 100$$

(1)

where $W\_nanocellulose$ is the freeze-dried nanocellulose weight (g) and $W\_cellulose$ is the initial cellulose weight (g).

4.3. Characterizations of Nanocellulose

Morphological characterization was performed using a field emission scanning electron microscope (FE-SEM) (JEOL JSM-IT800, Peabody, MA, USA) operated at 20 kV with a magnification of 30,000×. A nanocellulose suspension was prepared by dispersing the sample in distilled water, and then it was stirred for 30 min. Subsequently, 10 µL of the dispersion was dropped onto a copper grid and allowed to air-dry under light for 20 min. The length and width of the nanocellulose fibrils were measured using ImageJ software (version 1.54g; National Institutes of Health, Bethesda, MD, USA).

The color change in cellulose and nanocellulose samples was evaluated to determine the total color difference (TCD) using a colorimeter (HunterLab ColorFlex EZ 45–0, Hong Kong, China) based on the CIE Lab* system. The measured L*, a*, and b* values were used to calculate the total color difference (ΔE*), whiteness index (WI), and yellowness index (YI). The total color difference was calculated according to the CIE76 equation [29]:

$${\Delta E}^{\ast }=\surd ({(L_t^{\ast }- L_0^{\ast })}^2+{(a_t^{\ast }- a_0^{\ast })}^2+{\left(b_t^{\ast }- b_0^{\ast }\right)}^2)$$

(2)

The WI was calculated according to ASTM E313 [30]:

$$WI=100-\surd ({\left(100-L^{\ast }\right)}^2+{{(a}^{\ast })}^2+{{(b}^{\ast })}^2)$$

(3)

The YI was determined using:

$$YI=142.86\times b^{\ast }/L^{\ast }$$

(4)

4.4. Chemical Characterization of Nanocellulose

FTIR spectra (JASCO, Pfungstadt, Germany) were used to evaluate the functional groups of nanocellulose powder from bagasse and durian rind. Dry nanocellulose in a quantity of 0.002 g was pressed into a pellet with KBr. The transmission level was measured in the wavenumber range of 500–4000 cm⁻¹.

4.5. In Vitro Test of Nanocellulose Adsorption Capacity Against Mycotoxin

The evaluation of mycotoxin adsorption capacity was conducted following the protocol established by Tapingkae et al. [31]. Ochratoxin A (OTA) was prepared in a concentration of 10 ng/mL to determine the optimal concentration for standard testing with nanocellulose. For the experiment, bagasse-derived nanocellulose was prepared at concentrations of 33, 3, 0.3, and 0 mg/mL, while durian rind-derived nanocellulose was prepared at 36.5, 3.65, 0.36, and 0 mg/mL. A 100 µL aliquot of each nanocellulose suspension was pipetted into 1.5 mL microtubes. Subsequently, 100 µL of each OTA concentration was added to the respective microtubes containing nanocellulose. The mixtures were incubated for 30, 60 and 180 min to allow for sufficient time for the toxin and nanocellulose to reach adsorption equilibrium. After incubation, the samples were subjected to centrifugation at 15,000× g for 15 min. The resulting supernatant was carefully collected and set aside for further analysis using ELISA (R-Biopharm, Darmstadt, Germany). For the ELISA procedure, microwells were organized in a designated container to accommodate both standards and test samples. The locations and identities of all samples were meticulously recorded. Fifty microliters of the standard solutions and sample supernatants were dispensed into their respective wells, using new pipette tips for each standard preparation. An additional 50 µL of conjugate reagent was added to each well, and the plate was incubated at room temperature and shielded from light for five minutes. At the end of the incubation period, the liquid was decanted from the wells. The plate holder was inverted and tapped firmly against absorbent paper to ensure complete drainage. Each well was then rinsed with 250 µL of washing buffer, and the buffer was discarded as previously described. This washing process was repeated three times in total. Next, 100 µL of substrate solution was added to each well and mixed gently by manually agitating the plate. The plate was incubated for three minutes at ambient temperature away from light. Finally, 100 µL of stop solution was added to each well, followed by moderate agitation. Immediately after, absorbance was measured at 450 nm and recorded without delay.

4.6. Mycotoxin Adsorption Capacity of Nanocellulose via Chicken In Vitro Intestinal Digestion Model

The mycotoxin adsorption capacity of nanocellulose was evaluated using the protocol described by Tapingkae et al. [30], which simulates the avian digestive system. Contaminated feed samples (3 g) were transferred into 50 mL sterile tubes, followed by the addition of nanocellulose (50 mg). The sample were mixed thoroughly and incubated for 0, 3, 6, and 12 h. After that, 0.03 M hydrochloric acid (HCl, 10 mL) was added and the samples were thoroughly mixed. The pH was adjusted to 5.2 using 0.2 M sodium hydroxide (NaOH). The samples were subsequently incubated at 40 °C with continuous stirring at 19 rpm for 30 min to simulate crop digestion. Following the crop digestion phase, pepsin enzyme (3000 U/g feed) and 1.5 M hydrochloric acid (2.5 mL) were added to the samples. The pH was adjusted to the range of 1.4–2.0 to simulate gastric conditions. The tubes were incubated at 40 °C with continuous stirring at 19 rpm for 45 min. Upon completion of the gastric digestion phase, 8× pancreatin (6.84 mg/g feed) and 1.0 M sodium bicarbonate solution (6.5 mL) were added to each tube. During this phase, the pH of the samples was adjusted to the range of 6.4–6.8 to simulate intestinal conditions. The samples were subsequently incubated for an additional 2 h under the same stirring conditions (40 °C, 19 rpm). Following complete digestion simulation, the samples were centrifuged at 2000× g for 30 min. The supernatant was collected, filtered through a 0.45 µm filter membrane, and stored at −20 °C until analysis. Mycotoxin concentration in the samples was subsequently determined using enzyme-linked immunosorbent assay (ELISA).

4.7. Application of Nanocellulose on Adsorption Mycotoxin Contaminate in Poultry Farm Diets

Feed samples were collected from nine commercial broiler farming in Chiang Mai Province, Thailand. Feed samples were initially analyzed for mycotoxin contamination using ELISA, quantifying three analytes: ochratoxin A (OTA), fumonisin (FM), and aflatoxin (AF). Bagasse-derived nanocellulose and durian rind-derived nanocellulose were prepared at concentrations of 33 mg/mL and 36.5 mg/mL, respectively. A 100 µL aliquot of each nanocellulose suspension was pipetted into 1.5 mL microtubes. Subsequently, 100 µL of each mycotoxin concentration was added to the respective microtubes containing nanocellulose. The mixtures were incubated for 180 min. After incubation, the samples were subjected to centrifugation at 15,000× g for 15 min. The resulting supernatant was carefully collected and set aside for further analysis using ELISA.

The percentage of adsorbed mycotoxins was calculated using the following equation:

$$\% \mathrm{a}\mathrm{d}\mathrm{s}\mathrm{o}\mathrm{r}\mathrm{p}\mathrm{t}\mathrm{i}\mathrm{o}\mathrm{n}={\displaystyle \frac{C_{ads}}{C_0}}\times 100= {\displaystyle \frac{(C_0-C_{eq})}{C_0}} \times 100$$

where C_ads is the concentration of adsorbed mycotoxins, C₀ is the concentration of mycotoxins in the supernatant of the control sample without nanocellulose (no adsorbent), and C_eq is the residual mycotoxin concentration in the supernatant after incubation with nanocellulose. C_ads is calculated as C_ads = C₀ − C_eq.

4.8. Statistical Analysis

The averages and standard deviations for each data set have been computed. Utilizing a critical p-value of 0.05, a one-way ANOVA was employed to assess the significance of differences. Statistical analysis was conducted using SPSS Statistics (IBM SPSS Statistics 30.0, New York, NY, USA).