Dietary effects of cottonseed and vitamin E on greenhouse gas emissions from cattle feces analyzed in biodigesters

Abstract The objective of this work was to evaluate the effect of diets with cottonseed and vitamin E on the greenhouse gas (GHG) emissions of cattle feces analyzed in biodigesters. Animal feces were evaluated in 18 laboratory-scale anaerobic digesters through the following treatments: control, feces from cows fed with a control diet; CS, feces from cows fed with a cottonseed diet; and CSVitE, feces from cows fed with a cottonseed and vitamin E diet. The production of biogas and GHG was measured, and no differences were observed for the emissions of biogas, methane (CH4), and nitrous oxide (N2O). The partition of the gross energy of the anaerobic digestion process was similar among treatments. On average, 28% of gross energy fed was released as CH4, 47% during digestion, and 25% as other gases and heat. The addition of cottonseeds to the cattle diet increases the concentration of CH4 and reduces that of CO2, but it does not affect the total production of CH4, CO2, and N2O in the biodigesters. The inclusion of vitamin E in the diet has no effect on GHG production in the biodigesters.


Introduction
The manipulation of rumen fermentation by supplementing lipids in cattle diets (Wanapat et al., 2015) is a strategy to reduce methane (CH 4 ) emissions by ruminants.Lipids are directly related to fatty acids and, therefore, may be increased in diets through the addition of ingredients such as cottonseed (CS), a byproduct of the cotton fiber industry extensively used as a source of fiber, protein, or fat in cattle nutrition (Warner et al., 2020), due to its high concentrations of unsaturated fatty acids (Paim et al., 2014).Vitamin E has been used in animal feed to inhibit lipid oxidation (Juárez et al., 2012), relieving the effects caused by free radicals, being beneficial to rumen fermentation as it improves nutrient digestibility (Vázquez-Añón & Jenkins, 2007).
Some studies have investigated the addition of whole cottonseed and vitamin E in cattle diets (Polviset et al., 2015;Ferrinho et al., 2018;Nogueira et al., 2019Nogueira et al., , 2020)).However, there are no known works associating these ingredients with the impact of feces on the environment or potential use in digesters.According to Møller et al. (2014), the effects of changes in cattle's diets on biogas and GHG emissions from feces need to be further studied.
Biogas is a mixture of CH 4 and carbon dioxide (CO 2 ) with some trace gases (Mata-Alvarez et al., 2014), resulting from the conversion of organic substrates during the biological process of anaerobic digestion (Zhang et al., 2016).Biogas and CH 4 production from organic matter, defined as volatile solids, have been evaluated during the testing of anaerobic digestion under controlled or monitored conditions in laboratory batch on a small scale (Kunz et al., 2022).Nogueira et al. (2020), for instance, observed that the inclusion of cottonseed in cattle's diet improved rumen fermentation and reduced rumen CH 4 production.
However, although different feeding strategies can reduce enteric CH 4 emission from stored manure, or improve ruminal conditions (Benchaar & Hassanat, 2019), the subsequent effects on GHG emissions from feces still remain unclear.
The objective of this work was to evaluate the effect of diets with cottonseed and vitamin E on the GHG emissions of cattle feces analyzed in biodigesters.

Materials and Methods
The study was conducted in the summer of 2017 at Universidade de São Paulo, located in the municipality of Pirassununga, in the state of São Paulo, Brazil (21°59'45"S, 47°25'37"W, at 625 m above sea level).The climate of the season is classified as Aw according to the Köppen-Geiger, with an average temperature of 21.7°C and an average annual rainfall of 1,346 mm.The experiment was approved by the ethics committee on animal use of Universidade de São Paulo (application number 0092013).
Six non-pregnant and non-lactating Holstein dairy cows, with an average body weight of 876±16.1 kg, were arranged in individual pens with free access to water and sand bedding.The cows were fed ad libitum twice a day at 8:00 a.m. and 4:00 p.m with a highenergy ration containing 87% of concentrates, using sugarcane bagasse as roughage.Vitamin E (Lutavit E 50, BASF Australia Limited, Victoria, Australia), with 50% alpha-tocopheryl acetate, was included at an amount of 500 IU per day.Three dietary treatments were tested: control, control diet; CS, control diet with 30% cottonseed replacing ground corn grain; and CSVitE, CS diet supplemented with vitamin E. The ingredients and chemical compositions of the diets are presented in Table 1.
For feeding and feces collection, cows were arranged in a 3x3 Latin-square design with three experimental periods, each lasting 21 days.The first 14 days of each period were used for diet adaptation.Feces were collected during the next 15 to 21 days, twice a day, after the cows were fed.Representative samples -feces of each animal in each period -were pooled together for testing.The substrates (feces and water) were prepared to ensure an estimation of 5% of total solids.The feces were loaded into 18 laboratoryscale digesters, consisting of a 75 mm reactor, a 100 mm gasometer, and a 150 mm digester made with three PVC pipes, adapted from Sunada et al. (2018).
In a chamber with controlled temperature, the anaerobic digestion test was carried out under mesophilic conditions, ideal for digestion kinetics (Metcalf & Eddy, 2014).The digesters were arranged in a completely randomized design with three feces treatments (control, CS, and CSVitE) and six replicates.The digesters were started up using the feces with no inoculum.Every 15 days, biogas volume was calculated, through the vertical displacement of the gasometer, in centimeters, and the obtained value was standardized for the conditions of 1.0 atm and 20ºC.
Every time biogas volume was measured, biogas samples were collected with a syringe connected to the gas log on top of the gasometer.The concentrations of CH 4 , CO 2 and N 2 O were determined using the Trace 1300 gas chromatograph (Thermo Fisher Scientific, Milan, Italy), equipped with a flame ionization detector at 280°C, with a 3.5 m column packed with Porapak N matrix (Merck, Darmstadt, Germany), according to Kamiński et al. (2003).The volumes of CH 4 , CO 2 , and N 2 O were calculated by multiplying biogas volume by its respective concentration.Specific gas yield (per gram of volatile solids fed or destroyed) was calculated by dividing the total gas production by the amount of volatile solids fed (before anaerobic digestion) or destroyed (difference between volatile solids fed and eliminated).The test was considered finished when biogas production ceased.The nutrients fed (NF) and eliminated (NE) were then weighed to calculate dry matter (DM) content in grams.
Other gases and heat released, expressed in megajoule, were determined as follows: gross energy fed minus energy released as CH 4 minus gross energy eliminated in the digestate.The percentage of gross energy released, when expressed as percentage of gross energy fed, was calculated by division between gross energy of CH 4 or other gas and heat or digestate and gross energy fed and multiplied by 100.
Individual feed and feces samples, before and after anaerobic digestion, were collected and composited in representative samples on an equal-weight basis.The samples were dried in a forced-air oven, at 60°C, for 48 hours, and then ground to pass a 1.0 mm Wiley mill screen and analyzed.DM content was determined by method 930.15 of Association of Official Analytical Chemists (AOAC) (Cunniff, 1995) in the forced-air oven at 105°C for 2 hours, followed by cold weighing.Nitrogen content was obtained by the micro Kjeldahl method, being multiplied by 6.25 to calculate crude protein (Cunniff, 1995).Ether extract was obtained using light petroleum ether in the Soxhlet extraction apparatus, as described in method 920.39 (Cunniff, 1995).Neutral detergent fiber, acid detergent fiber, and lignin were determined by the methods described in the literature (Van Soest et al., 1991), using the Filter Bag Technology (Ankom Technology, Macedon, NY, USA) and heat-stable α-amylase as in method 973.18 (Cunniff, 1995).The levels of total solid and volatile solids were measured according to American Public Health Association (APHA) (Rice et al., 2012).
The experimental design was completely randomized, with 18 experimental units (digesters) for the three treatments (control, CS, and CSVitE) and six replicates.Data were tested for residual normality using the Shapiro-Wilk test.Statistical analyses were performed with the SAS, version 9.3, software (SAS Institute Inc., Cary, NC, USA) using mixed model, with the fixed effect of treatment.The analyses included descriptive statistics, in which mean values and standard errors of the mean were calculated.The variables were analyzed using the following model: Y ijkl = μ + T i + e ijk , where Y ijkl is the dependent response variable, μ is the overall mean, T i is treatment effect, and e ij is the residual error.
Contrast statements, at p≤0.05, were used to evaluate differences between the following treatments: control vs. CS and CSVitE; and CS vs. CSVitE.
Methane yield curve parameters were estimated from the methane yield of individual digesters using the Gompertz model (Kafle & Chen, 2016), specifically through the following three equations: y 1 =A exp[-B exp(-k)] where y1 is methane yield at anaerobic digestion days, A is asymptotic methane yield, B is the interaction constant, k is the yield constant rate, and exp is the base of natural logarithm 2.7183.
t 1 = ln B / exp where t 1 is the point of inflection, ln is the logarithmic, B is the interaction constant, and k is yield constant rate.
y 1 = A / K where y 1 is methane yield at inflection point, A is asymptotic methane yield, and exp is the base of natural logarithm 2.7183.

Results and Discussion
Feces composition differed in accordance with the diet (Table 2).Cottonseed treatments had a higher concentration of total solids (TS), volatile solids (VS), neutral detergent fiber (NDF), acid detergent fiber (ADF), and lignin when compared to the control.This result is indicative that the digesters were not able to destroy the highest content of TS and VS present in the feces from cows fed with the diets with cottonseed, as confirmed by the similar values of TS and VS destroyed.Consequently, there was a higher elimination of TS and VS in the cottonseed treatments.
No significant differences were observed between TS destroyed and VS destroyed, whose mean values were 49.6% and 53.7%, respectively.Volatile solids destroyed are commonly used to measure the performance of processes, being a direct indicator of the metabolic activity of a microorganism community (Møller et al., 2004).The destruction of VS in cattle manure in the anaerobic digestion process is typically in the range of 30-45% (Davidsson et al., 2008).In the present study, regardless of the treatment, the destruction of TS and VS was over 45% and was indicative of a sludge stabilization process and good reactor performance in the batch digesters.As a result, feces from cattle fed with cottonseed, with or without vitamin E, did not affect biogas production.
The microorganisms present in the control degraded more protein and organic matter than those in CS and CSVitE (Table 2).These results suggest that the microorganisms in the control had to hydrolyze higher quantities of crude protein and organic matter to obtain nutrients, as reported by Coelho et al. (2022).By contrast, the microorganisms in the CS and CSVitE had to hydrolyze higher quantities of protein from the NDF provided by the cottonseeds included for nutrients.This means that both treatments showed a similar destruction of VS and TS, in addition to CH 4 yield.Although the CS and CSVitE have a higher amount of TS available for the anaerobic digestion, the nutrients  2012) concluded that a higher concentration of fiber in the diet may favor a lower reduction in TS and VS, whereas Kafle & Chen (2016) observed a strong negative relationship, with a coefficient of determination (R 2 ) of 0.85, between lignin concentration in manure and CH 4 emissions.Therefore, although there was an increase in lignin content in the feces in the cottonseed treatments, there was no reduction in CH 4 production.
The crude protein content of a diet is the major factor determining the percentage of excreted N (Külling et al., 2002).In the present study, a similar crude protein content was obtained for the diet and substracts used to load the digesters (Table 2).Since the crude protein content in the substracts is a N source for the process of N 2 O production, a similar crude protein resulted in similar N 2 O emissions.The same authors reported similar N losses and N 2 O emissions for dairy manure under anaerobic conditions.
Regarding biogas, methane concentration was higher in the cottonseed treatments when compared to the control (Table 3).However, the diets with cottonseeds had a lower CO 2 concentration, CO 2 yield per gram of VS fed, and CO 2 yield per gram of VS destroyed, with no significant differences for the N 2 O concentration in the biogas, total yield, yield, and Gompertz parameters.Mathot et al. (2012) only found alterations in feces characteristics but not in GHG emissions, as observed here.
To calculate the CH 4 emissions of the manure of dairy cows, the Intergovernmental Panel on Climate Change (IPCC) (Dong et al., 2006) suggests a default value of 0.240±36 L g -1 of VS as the maximum amount of CH 4 producing capacity of manure.In the present work, the specific CH 4 productivity, measured in terms of VS destroyed, was, on average, 0.303 L g -1 of VS destroyed, which is indicative of the complete degradation of the organic components of the feces.The ultimate CH 4 yield, obtained in terms of VS fed (Dong et al., 2006), was 54.0% of the theoretical CH 4 yield, confirming the reduction of 53.7% in VS.The ultimate CH 4 yield will always be lower than the theoretical yield because a fraction of the feces is used to synthesize bacterial mass; therefore, the fraction of the organic material lost in the effluent and lignincontaining compounds will only be degraded to a limited degree (Franco et al., 2007).
The cumulative CH 4 yield was estimated by the Gompertz curve (Figure 1), which explains most of the variability in the response data, confirmed by the higher R 2 values obtained.A CH 4 yield difference of 3.4% was observed between the measured and predicted values, which is in alignment with Kafle & Chen (2016), who concluded that the Gompertz model was the one that better predicted CH 4 yield.Inflection point was reached, on average, on day 38 when cumulative CH 4 represented 38% of total CH 4 yield, which increased up to day 38, after which it decreased until day 220, when yield ceased.
The cottonseed treatments had higher gross energy fed and released in the digestate when compared to the control (Table 4).However, no differences were observed for gross energy released when expressed as a percentage of gross energy fed.On average, 47% of the gross energy fed was released during digestion, 25% as other gas and heat, and 28% as CH 4 that can be used for electricity or heat generation.

Conclusions
1.The addition of cottonseeds to cattle diets increases the concentration of CH 4 and reduces that of CO 2 , but does not affect the total yield of CH 4 , CO 2 and N 2 O in biodigesters.
2. The inclusion of vitamin E in cattle diets has no effect on greenhouse gas production in the biodigesters.

Figure 1 .
Figure 1.Cumulative CH 4 production in relation to anaerobic digestion time observed and adjusted by the Gompertz curve.The inflection point represents the day when the maximum point of CH 4 production occurred.Control, control diet; CS, control diet with 30% cottonseed replacing ground corn grain; and CSVitE, CS diet supplemented with vitamin E.

Table 1 .
Ingredients and chemical composition of dietary treatments for dairy cows.

Table 3 .
Biogas and greenhouse gas emissions of batch anaerobic digesters loaded with feces from dairy cows fed with different diets.Control, control diet; CS, control diet with 30 % cottonseed replacing ground corn grain; and CSVitE, CS diet supplemented with vitamin E. (3) SEM, standard error of the mean.(4)Contraststatements were used to evaluate differences between the following treatments: C1, control vs. CS and CSVitE; and C2, CS vs. CSVitE.*Significant at 5% probability.ns Nonsignificant.

Table 4 .
Gross energy released and eliminated by batch anaerobic digesters loaded with feces from dairy cows fed with different diets.Control, control diet; CS, control diet with 30% cottonseed replacing ground corn grain; and CSVitE, CS diet supplemented with vitamin E.(2) SEM, standard error of the mean.(3)Contraststatements were used to evaluate differences between the following treatments: C1, control vs. CS and CSVitE; and C2, CS vs. CSVitE.Statistical significance was declared at p≤0.05.*Significant at 5% probability.ns Nonsignificant.