Biosolids carbohydrate characterization

Five PB from different sources (PB1, PB2, PB3a, PB3b) including a DPS were chosen from previous studies conducted in eastern Canada (Gagnon et al. 2010, 2012; Gagnon and Ziadi 2012). Each material was kept frozen at −18 °C since its initial experimental use and thawed just before carbohydrate analysis and lab setting. Briefly, PB1 was produced from thermomechanical pulp, PB2 from bleached Kraft pulp with some municipal effluent, and PB3a and PB3b from acid-treated bleached Kraft pulp produced at a different time. All mills used either activated sludge (PB1, PB2, and DPS) or sequential aerobic biological reactor (PB3a and PB3b) as secondary wastewater treatment (Environnement Québec 2020). Chemical properties of the PB are given in Table 1. We expected that freezing would not cause a transformation of material composition as opposed to oven-drying samples (Ajiboye et al. 2004).

Table 1.

Main chemical properties of combined paper mill biosolids (PB) and deinking paper sludge (DPS) used in incubation and in the field (all data on dry matter basis except moisture).

Total neutral carbohydrates in PB were determined in triplicate according to the procedure of Chantigny and Angers (2008). Briefly, 0.10 g (±0.001) of finely ground (<0.15 mm) material dried at 30 °C was presoaked with 8 mL of 12 mol·L⁻¹ H₂SO₄ in 250 mL centrifuge bottles, gently mixed and left to stand on the bench for 2 h. Deionized water (184 mL) was then added to dilute the acid to 0.5 mol·L⁻¹. After incubation at 85 °C for 24 h, the slurries were cooled to room temperature and centrifuged for 10 min at 10 000 r·min⁻¹. The supernatant (20 mL) was decanted into a plastic vial and stored at −20 °C until analysis. Prior to analysis, samples were neutralized to pH 6.5–7.0 with 0.5 mol·L⁻¹ NaOH, filtered on Whatman #934-AH, and then successively purified through a solid phase anion-exchange column and a solid phase cation-exchange column (Supelclean LC-SAX and LC-SCX, respectively; Supelco Inc., Bellefonte, PA, USA). The purified extracts were quantified for neutral carbohydrates (glucose, galactose, mannose, arabinose, xylose, fucose, and rhamnose) using high-performance anion-exchange chromatography with pulsed amperometry detection (Dionex model ICS-5000, Sunnyvale, CA, USA).

Soluble neutral carbohydrates were also extracted in triplicate using the hot water procedure (Chantigny and Angers 2008). This fraction contains the most labile carbohydrates. In this case, 0.10 g (±0.001) of finely ground (<0.15 mm) material was incubated with 30 mL distilled water in 50 mL centrifuge tubes at 85 °C for 24 h. After the incubation, the slurries were cooled to room temperature and centrifuged for 10 min at 10 000 r·min⁻¹. Each hydrolysate was then transferred to new tubes with 1 part of 12 mol·L⁻¹ H₂SO₄ to 23 parts of supernatant, incubated again at 85 °C for 24 h, cooled to room temperature and centrifuged. The supernatant (20 mL) was decanted into a plastic vial and stored at −20 °C until analysis. The procedures of neutralization, purification, and quantification followed the same as described for total neutral carbohydrates.

In addition to carbohydrates, the content in hemicellulose-, cellulose-, and lignin-like fractions was also determined on dried, ground PB samples (<1 mm) using the method of Van Soest (1963) as reported in Robin (1997). Hemicellulose is a complex carbohydrate polymer composed of xylose, mannose, galactose, glucose, and arabinose, whereas cellulose is essentially made up of glucose (Pérez et al. 2002). This step characterizes organic matter of materials, which can allow to predict their dynamic and stability once incorporated into the soil (Abiven et al. 2008). Extraction was done with the Ankom Filter bag technique for the determination of NDF (Ankom technology 2017a), ADF (Ankom technology 2017b), and ADL (Ankom technology 2020). The different fractions are obtained as follows:

(1)

(2)

(3)

Field study

The field trial was conducted over 9 yr, from 2000 to 2008 at Yamachiche (lat. 46°17'N, long. 72°48'W), QC, Canada, on a Chaloupe loam (Table 2) to evaluate the impact of PB applied alone or with liming by-products on crop yield and soil chemical properties (Gagnon and Ziadi 2012). All treatments were surface applied every year, at post-seeding of annual row crops (grain corn in 2000–2002, 2004, and 2006–2008, dry bean (Phaseolus vulgaris L.) in 2003, and soybean in 2005). The PB along with crop residues was incorporated (depth of 5–7 cm) before the following season by disk harrowing.

Table 2.

Selected characteristics of soils used in the field and in incubation studies.

For this study, only treatments with an application rate of 0, 30, and 60 Mg PB·ha⁻¹ (wet weight) were considered. The 30 Mg fresh material·ha⁻¹ corresponded to the typical annual rate applied on cropland in Québec (Hébert 2016), and those reported in Ontario (15–30 Mg dry matter·ha⁻¹; Flemming et al. 2017). Supplemental inorganic N as Ca-NH₄NO₃ (45 kg N·ha⁻¹ averaged over the years) was, however, added to the 30 Mg PB·ha⁻¹ rate in 2004 and 2006–2008 to ensure adequate corn N nutrition. Plot size for each experimental unit was 3 m × 10 m, and experimental treatments were distributed within a randomized complete block design with four replicates. Among the five materials retained for carbohydrate characterization, PB1 was notably used in the 2000–2002 period, and PB3a was used in 2003–2008 (Table 1). The different choice was related to the PB availability from neighborhood mills and to a change for a more fertilizing material from 2003.

Soils were collected at random from each selected plot with a spade (composite of three cores) from the 0 to 10 cm layer after 3, 6, and 9 yr of PB application and sieved in the field to <4 mm. Samples were taken before dry bean seeding in spring 2003 and immediately after harvesting soybean and corn in October 2005 and 2008, respectively. Soil samples were air-dried, sieved to pass a 2 mm screen and then ground to <0.15 mm. The strong acid soil hydrolysis was performed as described for biosolids characterization except that 2.00 g (±0.001) was used in the extraction.

Water-stable aggregation was measured by wet sieving 10 g of field-moist soil through a series of sieves with decreasing openings (2, 1, 0.5, and 0.25 mm) by total immersion in water and gentle shaking for 10 min (Angers et al. 2008). The aggregate size classes were defined as >2 mm, 1–2 mm, 0.5–1 mm, 0.25–0.5 mm, and <0.25 mm and were expressed as a percentage of total initial soil (oven-dry basis) after correction for sand content. The mean weight diameter (mm) was calculated as:

where X_i is the mean diameter of each aggregate size class (mm) and W_i is the proportion of the total sample weight recovered in the corresponding size class.

Incubation study

Samples used were from a previous laboratory study evaluating the N and P release from several PB under controlled conditions (Gagnon and Ziadi 2021). Two texturally different soils were collected from fields located near Québec, QC, Canada (lat. 46°41-47′N, long. 71°08-28′W): a Kamouraska clay from a spring wheat field and a St. Antoine sandy loam from a field under alfalfa (Medicago sativa L.)/timothy (Phleum pratense L.) pasture (Table 2). Six treatments were used, which included the five materials (PB1, PB2, PB3a, PB3b, DPS) applied at 50 Mg wet weight·ha⁻¹ and an unfertilized control (Table 1). We chose this rate because it could be applied to several crops without supplementary N (Gagnon and Ziadi 2021) and then allowed to better assess carbohydrate contribution from PB. Incubation took place for 16 wk in the dark in a controlled-environment chamber at 24.5 °C, 60% water-filled pore space, with periodic aeration using 500 mL glass jars.

Soils of all jars were collected at the end of incubation and treated in the same manner as those from the field study to determine the concentrations in total neutral carbohydrates. In addition, the hot-water soil hydrolysis was also performed as described for biosolids characterization with 2.00 g (±0.001) subsamples to determine the soluble neutral carbohydrates.

Statistical analysis

All data for soil studies were subjected to a Bartlett’s test, and no transformation was necessary. Neutral carbohydrates were analyzed as a sum, as individuals or grouped as being from microbe- (galactose plus mannose) or plant (arabinose plus xylose)-derived origin (Oades 1984). Data on fucose and rhamnose were not reported in the soil analysis due to their lack of significance. Statistical significance was set at p ≤ 0.05.

Material carbohydrates (individual and the sum) were submitted to a cluster analysis using the Ward’s minimum variance method. Using cluster as a procedure of statistical analysis allows to group materials similar in their composition and to relate this to the carbohydrate increases found in soil during incubation, even if properties of PB may vary from mill to mill. Results of clustering were subsequently submitted to a generalized linear model analysis of variance (SAS Institute 2010) to identify characteristics differing significantly among clusters. Total carbohydrates and soluble carbohydrates were analyzed separately. Other related variables not used in the cluster analysis, such as calculation of microbe- and plant-derived carbohydrates, percentage of total carbohydrates in total C, percentage of soluble carbohydrates in total carbohydrates and total C, and organic matter fractions (hemicellulose, cellulose, lignin), were also compared for each cluster.

Data from the field study were analyzed with treatments and years as fixed effects and replications as random effects, using the MIXED procedure (SAS Institute 2010) with a banded main diagonal covariance structure for carbohydrate composition and a compound symmetry covariance structure for aggregates. Main treatment effects were compared using the least significant difference (5%). Correlation analysis was used to establish links between soil total neutral carbohydrates and water-stable aggregation.

Treatment effects in the incubation study were evaluated as a factorial PB × soil using the MIXED procedure (SAS Institute 2010). Treatment means were compared using the least significant difference (5%) and single degree-of-freedom contrasts for comparison of material clusters between them and with the unamended control, and within each cluster. Regression analysis was performed to relate the content of carbohydrates in materials to that found in the amended soils at the end of incubation.

Biosolids carbohydrate characterization

Cluster analysis identified two groups of PB for total carbohydrates. The first group comprised PB1, PB3a, and PB3b, and the second group PB2 and DPS (Fig. 1A). The first cluster separated from the other by a significantly higher content in galactose, mannose, and arabinose and a lower content in xylose (trend, p = 0.07), while having the same amounts of carbohydrates as a sum (Table 3). As a result, the carbohydrates of microbial origin (galactose and mannose) were also higher in the first group.

Fig. 1.

Dendrogram for clustering of (A) total neutral carbohydrates and (B) soluble neutral carbohydrates of combined paper mill biosolids (PB) and deinking paper sludge (DPS). Material description is defined in Table 1.

Table 3.

Characterization of paper mill biosolids in total neutral carbohydrates according to clustering analysis.

Glucose was the most abundant carbohydrate in the total composition (from 63% to 80%; Table 3), and with mannose and xylose formed 90% of total carbohydrates for cluster 1 and 97% for cluster 2 (p = 0.010). McGovern et al. (1983) reported that the composition of PB in total carbohydrates was quite constant between Wisconsin mills of Kraft, mechanical or recycled process, averaging about 65% glucose, 25% xylose, and 10% mannose. Our results agree for glucose, but a difference was noted for xylose and mannose that cannot be explained by the pulping process or wastewater treatment themselves. Possible explanation could be the tree species involved in the pulp and paper production. Softwood conifers (Picea and Abies spp.) are the main source of wood in the pulp and paper mills in the province of Quebec. However, hardwood such as quaking aspen (Populus tremuloides Michx.) may enter the paper-making chain depending on desired finished end-product, local wood supply, and type of recycled fibers. These species are composed of a larger proportion of xylose and a lower proportion of mannose and galactose than softwoods (Hakkila 1989).

Cluster analysis identified two groups of PB for water-soluble carbohydrates. The grouping, however, differed slightly from that for total carbohydrates. The first group comprised PB2, PB3a, and PB3b, and the second group PB1 and DPS (Fig. 1B). The separation of clusters was significant only for the sum of carbohydrates, with a trend (p = 0.05–0.07) for galactose and mannose, with higher values in cluster 1 (Table 4). The water-soluble carbohydrates formed between 1.7% and 4.9% of total carbohydrates but between 4% and 35% for those derived from microbes, including rhamnose and fucose. The carbohydrate fraction extractable with hot water has been found to be enriched in exocellular microbial polysaccharides (Haynes and Francis 1993; Gunina and Kuzyakov 2015). Results indicated that, contrarily to total fraction, the water-soluble fraction was composed of carbohydrates of various forms.

Table 4.

Characterization of paper mill biosolids in soluble neutral carbohydrates according to clustering analysis.

In this study, there was a trend (p = 0.08) for a lignin content twice as high in cluster 1 than in cluster 2, while the composition in hemicellulose and cellulose was similar (Table 3). Lignin, due to its polyphenolic complex structure, is recalcitrant to microbial decomposition (Bahri et al. 2008) and is considered as a precursor of soil humic substances (Robin 1997). According to Abiven et al. (2008), lignin has the greatest impact on soil aggregate stability, while the hemicellulose plus cellulose fraction contributes to long-lasting effect. Materials with a high lignin content decompose more slowly and their influence on aggregation is low (N’Dayegamiye 2006). We can then expect that PB of cluster 2 (PB2, DPS) could have the most positive effect on soil aggregation, at least in short term. Chantigny et al. (2000) reported that 30% of initial total carbohydrates in DPS was rapidly decomposed and then contributed to the formation of large (>1 mm) water-stable aggregates (Chantigny et al. 1999).

Soil carbohydrates under field study

Concentrations of total neutral carbohydrates were strongly affected by the repeated applications of PB (Table 5). Concentrations in all carbohydrates except arabinose were significantly increased by the 30 Mg PB·ha⁻¹ rate. The application of 60 Mg PB·ha⁻¹ also increased the concentrations in arabinose, and gave higher concentrations in glucose and mannose than the 30 Mg PB·ha⁻¹. These two carbohydrates composed of around 80% of total carbohydrates in the applied materials (Table 3). Relative to unamended control, the 60 Mg PB·ha⁻¹ rate increased the concentrations of total neutral carbohydrates by 42% in spring 2003 and by 96% in fall 2005 and fall 2008 (Fig. 2), while the concentrations of individual carbohydrates were 1.6–2.7 times greater at the last sampling, in fall 2008 (data not shown). The ratio of galactose and mannose to arabinose and xylose was unaffected by treatments and varied from 0.9 to 1.6 depending on sampling dates (Table 5), meaning that both microbial and plant origin carbohydrates contributed equally to the increase in soil total carbohydrates.

Table 5.

Effect of repeated paper mill biosolids (PB) application in field on the concentrations of total neutral carbohydrates in the 0–10 cm layer of a loamy soil.

Fig. 2.

Effect of combined paper mill biosolids at 30 (PB30) and 60 (PB60) Mg wet weight·ha⁻¹ on the concentration of total neutral carbohydrates in the 0–10 cm layer of a loamy soil after 3, 6, and 9 yr of annual applications under field conditions. PB1 was used in 2000–2002 and PB3a in 2003–2008.

A positive effect of applying PB, including DPS, on soil carbohydrates was reported elsewhere under field studies. Bipfubusa et al. (2008) found a 29% increase in total carbohydrate content of soil large macroaggregates 2 yr after the last of three annual PB applications at 40 Mg wet weight·ha⁻¹. However, they did not evaluate the carbohydrate composition. Chantigny et al. (2000) reported increases in several measured carbohydrate fractions of DPS-amended soils with a predominant contribution of microbial carbohydrates. Repeated PB application, even at low rates (20 Mg wet weight·ha⁻¹), may sustain soil physical properties longer and overcome the transient binding nature of decomposing freshly added labile C (N’Dayegamiye 2006).

Soil aggregation under field study

The water aggregate stability was affected by PB application. The proportion of aggregates ≥1 mm increased with the 60 Mg PB·ha⁻¹ rate in the 0–10 cm soil layer over the course of the study (Table 6). Aggregate size between 1 and 5 mm is the most favorable for seedbed and plant growth (Braunack and Dexter 1989). At the opposite, repeated application of PB decreased the amount of microaggregates <0.25 mm while it did not impact the other size classes except 0.5–1 mm in fall 2008. On average, the mean weight diameter of aggregates increased from 1.76 mm in the untreated control to 1.92 mm in the plots receiving 60 Mg PB·ha⁻¹ annually.

Table 6.

Effect of repeated paper mill biosolids (PB) application in field on the water-stable aggregates in the 0–10 cm layer of a loamy soil.

In this study, the soil had a degraded physical structure and was under intensive annual cropping (Haynes et al. 1991), even though minimum tillage was practised. Addition of PB, due to their high cellulose and labile C content, promoted a high mineralization rate (Bipfubusa et al. 2005) which significantly improved structural stability over the years. N’Dayegamiye (2006, 2009) reported that organic materials such as PB with a low C/N ratio and a low recalcitrant-like lignin fraction enhanced soil aggregation after several applications but their benefit effects varied with soil type, being more consistent on silt and clay loams than on sandy loam. Other studies observed increases in ≥1 mm size class aggregates with cumulative PB application (Zibilske et al. 2000; Price and Voroney 2007). Abiven et al. (2009) classified different organic products according to their magnitude and time-to-maximum effect on aggregate stability and found that sludge category, which comprised PB and DPS, had a long-term effect (>3 mo) with a strong magnitude.

Components of organic C, such as carbohydrates, are partially involved in the stabilization of soil into aggregates (Angers and Mehuys 1989). Those of microbial origin appear to be the most effective and active (Oades 1984; Haynes and Francis 1993; Degens 1997; Puget et al. 1999), although an accumulation of microbial carbohydrates probably only indicates greater microbial decomposition of organic C (Degens 1997). In this study, significant correlations were found between total neutral carbohydrate content (sum and individual, 0.59 ≤ r ≤ 0.82) and proportion of large water-stable aggregates measured in fall 2005, after 6 yr of repeated PB application. Additional years of amendment continued to increase soil total carbohydrates (Fig. 2) without further impacting soil macroaggregation.

Soil carbohydrates under incubation study

The different PB induced a significant effect on the soil total neutral carbohydrate content at the end of the 16 wk incubation (Table 7), with a supply of between 2.4 (PB2) to 4.6 (PB3a) Mg total carbohydrate-C·ha⁻¹ at the application rate of 50 Mg wet weight·ha⁻¹. The Kamouraska clay soil had higher carbohydrate content as individual or sum than the St. Antoine sandy loam, but both soils responded similarly to PB application (no interaction of treatment × soil).

Table 7.

Effect of paper mill biosolids on the soil total neutral carbohydrates at the end of a 16 wk incubation.

All PB of cluster 1 (PB1, PB3a, PB3b) increased soil total neutral carbohydrates, whereas those of cluster 2 (PB2, DPS) did not produce any effect compared with the unamended control (Table 7). Carbohydrates in the amended soils may originate from PB or be the product of the microbial activity promoted by PB addition. With a low (<15) or overly wide (>40) C/N ratio, microbial polysaccharides production from organic amendments may be restricted due, respectively, to a deficiency of organic C in substrate or a low microbial activity (Avnimelech and Cohen 1989). Under different conditions, Chantigny et al. (2000) reported a strong increase in several fractions of soil carbohydrates in a field study the next summer when 50–100 Mg dry weight·ha⁻¹ of DPS (6.4–12.7 Mg total carbohydrate-C·ha⁻¹) were applied the preceding fall.

The increase in total carbohydrates for PB belonging to cluster 1 averaged 44% for the sum of carbohydrates, with 986 mg C·kg⁻¹ more glucose relative to the control (+70%), 282 mg C·kg⁻¹ more mannose (+45%), 200 mg C·kg⁻¹ more xylose (+43%), 140 mg C kg⁻¹ more galactose (+25%), and 79 mg C·kg⁻¹ more arabinose (+17%). Results from the incubation study confirmed those obtained in the field. It was in line with the composition of PB (Table 3) and agrees with the relative carbohydrate abundance in soil composition observed by Chantigny et al. (2000) with DPS. The response when applying PB differed from that with manure in which the contribution of each individual carbohydrate is diverse (Angers and N’Dayegamiye 1991; Xie et al. 2014).

The ratio of galactose and mannose to arabinose and xylose is indicative of the predominance of microbial carbohydrates (>2) or plant carbohydrates (<0.5; Oades 1984). In this study, the ratio was higher in the Kamouraska clay than in the St. Antoine sandy loam but was unaffected by treatments (Table 7), which was also observed in the field. This may indicate greater microbial decomposition of organic C in the clay soil (Degens 1997).

Contrary to soil total neutral carbohydrates, the soil water-soluble neutral carbohydrates, as a sum or individual composition, were little affected by treatments, likely due to high variability (Table 8). Only PB3a increased the soluble fraction relative to unamended control due to an increase in glucose and arabinose. This fraction extracted by hot water represents 5–10% of total fraction (Gunina and Kuzyakov 2015). However, it decomposes rapidly during the first few days after material application and microbial replenishment (Abiven et al. 2008) and was particularly involved in the stabilization of soil aggregates (Metzger et al. 1987; Haynes and Francis 1993; Degens 1997). Since this fraction is low and very sensitive to changes in soil management (Degens 1997), reapplication should be advised to see any positive effects on soil quality.

Table 8.

Effect of paper mill biosolids on the soil soluble neutral carbohydrates at the end of a 16 wk incubation.

Relationships between material carbohydrates and soil composition

The incubation study showed a close relationship between soil total neutral carbohydrate content and the amount of microbial (galactose + mannose) carbohydrates added by PB (r² = 0.83; Fig. 3A), which was not observed with the plant-derived (arabinose + xylose) (r² = 0.15) or total carbohydrates (r² = 0.22). In addition, the soil total microbial-derived carbohydrate content was also correlated with the amount of added galactose + mannose (r² = 0.82; Fig. 3B). Such links have never been reported before between material carbohydrates and soil composition once applied. With the positive link obtained in the field between the soil total (r² = 0.59) and microbial-derived (r² = 0.64) carbohydrates and the proportion of large water-stable aggregates, this indicated that the microbial carbohydrates would be an important contributor to the improvement of soil quality.

Fig. 3.

Relationship between amounts of microbial-derived carbohydrates applied and the concentrations of (A) total and (B) microbial neutral carbohydrates in the soil at the end of incubation (average over soil type). PB, combined paper mill biosolids; DPS, deinking paper sludge, applied at 50 Mg wet weight·ha⁻¹; Un, unamended control. Soil carbohydrates were converted in kg C·ha⁻¹ by using a soil bulk density of 1.2 g·cm⁻³ and an incorporation depth of 10 cm.

The soil water-soluble neutral carbohydrate fraction was relevant to soil aggregate stability, since this specific fraction generally associated with fungal development correlates better than the much larger acid-hydrolyzable fraction (Metzger et al. 1987). Despite a lack of significance for treatment effect, this fraction was closely related to the amount of total carbohydrates added (Fig. 4A) as well as the microbial-derived carbohydrates (Fig. 4B), plant-derived carbohydrates (Fig. 4C), and sum of soluble carbohydrates (Fig. 4D). This suggests that the carbohydrates in the total fraction replenished those in soluble fraction once they were degraded.

Fig. 4.

Relationship between amounts of (A) total, (B) microbial-derived, (C) plant-derived, and (D) soluble carbohydrates applied and the concentrations of water-soluble neutral carbohydrates in the soil at the end of incubation (average over soil type). PB, combined paper mill biosolids; DPS, deinking paper sludge, applied at 50 Mg wet weight·ha⁻¹; Un, unamended control. Soil carbohydrates were converted in kg C·ha⁻¹ by using a soil bulk density of 1.2 g·cm⁻³ and an incorporation depth of 10 cm.