Forecasting Potential Resources of Humic Substances in the Ukrainian Lignite

Abstract

The present research deals with forecasting the potential content of humic acids (HA) in Ukrainian lignite based on the coal’s physicochemical characteristics. The focus is on developing an experimental–statistical model that considers moisture content, volatile matter yield, and calorific value of lignite. The development of the humic acid yield’s dependence on some lignite parameters is based on both original research data and literature sources. Humic acids were extracted using alkaline solutions, and HA content was calculated for various lignite deposits in Ukraine. The adequacy check of the model showed a relative prediction error of up to 12%, indicating good agreement between the model and experimental data. The highest potential yield of humic acids was recorded for lignite from the Dnipropetrovsk region (Dnieper-Donets Basin), amounting to 53–56 wt.%. The presented results demonstrate the feasibility of using mathematical forecasting to assess the industrial potential of humic acid production from lignite.

1. Introduction

Lignite (also known as brown coal) is a sedimentary rock that belongs to one of the types of fossil coal [1]. It is formed through the prolonged decomposition of plant residues under the influence of high pressure, temperature, and microorganisms in conditions of limited oxygen access. Lignite is an intermediate stage between peat and hard coal in coalification. Lignite is primarily used as a fuel, which causes several issues:

-

The combustion of coal, including lignite, releases large amounts of carbon dioxide (CO₂), sulfur oxides (SO₂), nitrogen oxides (NO_x), and other harmful substances that contribute to climate change, acid rain, and air pollution [2,3,4,5];

-

Lignite has a lower calorific value than hard coal and other fuels, making it less efficient for energy production [6,7].

-

High moisture and ash content reduce its quality and complicate transportation and storage [8,9].

As noted in [10], the potential lignite reserves in Ukraine amount to approximately 5.4 billion tonnes. If only the recoverable (balance) reserves of lignite in Ukraine are considered, they are estimated to be about half that amount—approximately 2.3–2.6 billion tonnes (see

Table 1. Distribution of recoverable lignite reserves by regions of Ukraine as of 1 January 2024.

Region	Recoverable Reserves, kt		Number of Deposits
	A + B + C1	C2
Dnipropetrovsk Region (Ukrainian Shield, Central Uplift, Prydniprovsky Block)	1,033,945	258,053	18
Dnipropetrovsk Region (Dnieper-Donets Basin, Southern Flank)	286,699	0	3
Zhytomyr Region (Ukrainian Shield, Central Uplift, Volyn Block)	10,884	0	2
Zakarpattia Region (Transcarpathian Intermountain Basin, Chop–Mukachevo Zone)	38,745	0	4
Kirovohrad Region (Ukrainian Shield, Central Uplift, Kirovohrad Block)	750,833	39,604	44
Kharkiv Region (Dnieper-Donets Coal-Bearing Area)	389,985	0	1
Cherkasy Region (Vatutine Geological and Industrial District)	82,225	1524	8
Total:	2,593,316	299,181	80

) [10,11,12]. Meanwhile, according to the International Energy Agency (IEA), which bases its estimates on criteria of accessibility and economic feasibility, the proven reserves of lignite and sub-bituminous coal in Ukraine as of the end of 2024 were approximately 2.9 billion tonnes. This ranks Ukraine 15th in the world regarding such reserves [13].

However, new lignite deposits are currently not being explored in Ukraine, and existing deposits are practically not being exploited, despite their number and favorable location. This is primarily due to the impracticality or impossibility of using lignite in the energy industry’s traditional application sector.

On the other hand, lignite has potential for use in “green” (non-fuel) technologies. The authors are researching to develop innovative, environmentally safe technologies for the rational use of low-metamorphosed solid fossil fuel resources. It has been established that components of lignite (humic acids and/or humates) can be effectively used to improve certain performance characteristics of road bitumen [14], to produce advanced sorbents [15], and to develop biodegradable polymer materials with enhanced performance properties [16,17].

Humic acids (HA) were extracted from lignite for the abovementioned purposes. HA are amorphous polymeric mixtures with high molecular weight, typically black or brown in color, and are key components of soil organic matter and low-metamorphosed fossil fuels. Humic acids contain various active functional groups, including carboxyl, hydroxyl, and carbonyl groups [18,19]. The processes for extracting humic acids (HA) from lignite [20,21,22,23,24] generally involve alkaline extraction followed by precipitation of HA using strong mineral acids. At the same time, various bases and acids can be used. Additionally, the coal may be pretreated with various agents (such as sodium pyrophosphate, hydrogen peroxide, etc.) to partially depolymerize the organic fraction of lignite, thereby increasing the yield of the alkaline extract. However, the authors suggest that despite differences in HA extraction methods, the yield primarily depends on the characteristics of the raw material itself. For example, when using different reagents, the HA yield from one type of coal differed insignificantly, and vice versa—when using similar methods, but different lignite samples, the difference was more than three times [25,26]. Given that determining the HA content in lignite is a complex, time-consuming, and costly process, a decision was made to develop a mathematical model describing the HA yield as a function of lignite characteristics, based on both the authors’ research and the scientific literature. Such a model would enable analysis of HA content in Ukrainian lignite and support forecasting the feasibility of utilizing these deposits in the technologies developed by the authors.

2. Materials and Methods

2.1. Materials

To develop the model, the characteristics of lignite from the Mokrokalyhirske lignite open-pit (Cherkasy Region, Ukraine) were used (see

Table 2. Characteristics of lignite samples.

Deposit Name	Sample No.	Wa, wt.%	Aa, wt.%	Ad, wt.%	Va, wt.%	Vdaf, wt.%	Sdt, wt.%	Qsdaf, Kkal/kg	HAa, wt.%	HAdaf, wt.%	Elemental Composition, wt.%
											Cdaf	Hdaf	Sdaf	Ndaf	Odaf
Mokrokalyhirske, Ukraine	1	16.50	- **	47.90	-	57.80	2.11	-	-	87.58	80.65	4.39	3.52	1.25	10.19
	2	8.70	-	8.60	-	50.10	1.77	-	-	49.59	67.89	4.55	1.92	1.33	24.31
	3	30.40	-	36.40	-	69.70	2.72	-	-	82.86	60.58	4.82	3.64	1.29	29.67
	Average *	18.53	-	30.97	-	-	2.20	-	-	73.34	69.71	4.59	3.03	1.29	21.39
Chakwal, Pakistan	4	4.32	56.40	-	19.81	-	-	3236	25.50	-	26.73	12.39	5.05	0.00	55.83
	5	2.60	57.20	-	17.30	-	-	3021	27.40	-	20.03	0.37	0.01	0.00	79.59
	6	5.33	55.89	-	19.67	-	-	3074	25.30	-	30.38	2.75	18.57	0.00	48.30
	7	3.44	58.43	-	18.41	-	-	3381	25.40	-	31.01	5.27	38.26	0.00	25.43
	8	3.80	56.70	-	18.50	-	-	3256	27.10	-	14.70	1.69	13.04	0.00	70.57
	9	4.44	47.22	-	18.83	-	-	3188	28.30	-	22.38	17.46	0.00	0.00	60.16
Arenas de Rey and Puentes de García Rodríguez, Spain	10	9.60	5.40	-	46.40	-	-	-	-	47.90	60.91	4.15	6.35	0.88	27.71
	11	7.70	32.20	-	32.20	-	-	-	-	50.70	61.36	5.20	3.33	0.77	29.34
Thar, Pakistan	12	23.00	14.00	-	34.00	-	-	3809	35.00	-	-	-	-	-	-
Xiaolongtan, China	13	9.30	-	10.50	-	42.30	-	-	-	57.80	57.47	5.01	1.16	1.58	34.78

Note: * Average values of lignite parameters from the Mokrokalyhirske deposit, Ukraine.**—Data missing.

); this coal had previously been used by the authors for HA extraction in earlier studies [27]. Characteristics of foreign lignite samples were also employed to construct the mathematical model. These included coal from mines in the Chakwal region, Pakistan [28]; lignite from Arenas de Rey (1) and Puentes de García Rodríguez (2), Spain [29]; coal samples from Thar, Pakistan [30]; and Xiaolongtan lignite, China [31]. These samples provided HA yield data and coal property parameters required for modeling (see Table 2).

A relatively small data set was used to develop the model, which is because of the closure of lignite deposits in Ukraine and the small amount of literature data that is able to provide the necessary characteristics of brown coal and the yield of humic acids. However, the number of observations is 4.3 times greater than the number of selected parameters, and n > = m + 1, where n is the number of observations; m is the number of model regressors (n = 13, m = 3). Therefore, it can be argued that the number of observations is sufficient to unambiguously determine the parameters of the dependence of the yield of humic acids on the selected characteristics of lignite [32].

To verify the reproducibility and accuracy of the predicted humic acid yield based on the developed mathematical relationships, an additional coal sample was collected from the Morozivske deposit. The physicochemical characteristics and elemental composition of the sample are presented in

Table 3. Coal characteristics from the Morozivske deposit, Ukraine.

Wa, wt.%.	Ad, wt.%	Vdaf, wt.%	Elemental Composition, wt.%
			Cdaf	Hdaf	Ndaf	Sdaf	Odaf
26.40	13.00	61.80	69.55	4.65	0.70	3.37	21.73

.

The average characteristics of deposits from which Ukrainian coal was used for research are presented in

Table 4. Reference characteristics of coal from Ukrainian deposits.

Deposit Name	Wa, wt.%	Aa, wt.%	Ad, wt.%	Sdt, wt.%	Vdaf, wt.%	Qsdaf, MJ/kg	Qaf, MJ/kg
Mokrokalyhirske lignite open-pit mine	17.90	21.10	25.70	3.17	60.70	28.33	23.97
Morozivske deposit	14.00–15.20	15.00–20.20	17.60–23.50	2.70–4.00	61.10–68.70	24.62–28.68	18.56–24.48

, based on reference data [33].

By comparing the experimental average characteristics of samples from the Mokrokalyhirske lignite open-pit mine and the Morozivske deposit (Table 2 and Table 3) with reference data (Table 4), it can be concluded that the selected coal samples are informative and their properties correlate well with the reference data.

2.2. Methods

2.2.1. Extraction of Humic Acids from Lignite

Humic acids from the Ukrainian coal samples were obtained by treating lignite with an alkaline solution of sodium pyrophosphate, followed by extraction with a 1% sodium hydroxide solution and precipitation using an excess of hydrochloric acid. The extracted humic acids were separated from the solution by centrifugation, then washed and dried.

In [28], the International Humic Substances Society (IHSS) method was used. Humic acids were obtained by extracting lignite with KOH and precipitating with hydrochloric acid. The humic acid precipitates were separated by centrifugation, thoroughly washed with distilled water, and dried. In [29], HAs were obtained by extracting coal with NaOH. The suspension was centrifuged, and the supernatant was used for analysis. The HA content was measured colorimetrically. In [30], lignite was first treated with a hydrogen peroxide solution, then extracted with NaOH and precipitated using HCl. The solid humic acid was separated by centrifugation. In [31], HAs were extracted using a KOH solution; the extract was then treated with H₂SO₄ to adjust the pH to 1.0 for humic acid precipitation. The precipitate was separated by centrifugation, washed with water, and dried under vacuum.

2.2.2. Determination of Physicochemical and Technological Parameters

The determination of physicochemical and technological parameters for Ukrainian lignite was carried out using standardized methods: moisture content [34], ash content [35], volatile matter content [36], and elemental composition analysis [37]. The higher specific calorific value of lignite was calculated based on its elemental composition data [38]:

$$Q_s^d=0.3491·C^d+0.1783·H^d+0.1005·S^d-0.1034·O^d-0.0151·N^d-0.0211·A^d,$$

(1)

2.3. Development of a Mathematical Model and Determination of Humic Acid Yield Based on Lignite Characteristics

The STATISTICA software package was used to construct the equations of the specified functions.

To assess the adequacy of the obtained regression equations, the expected response function values (Y_i^reg) were calculated by substituting the experimental factor values (X₁–X₃) into the regression equations, based on which the residuals were determined:

∆Y_i = Y_i^reg − Y_i,

(2)

where Y_i—observed values obtained from experimental and literature data;

Y_i^reg—response function values calculated using the regression equations;

i—sample number.

The following indicators were calculated to analyze the constructed model. The formula for calculating the average relative approximation error is as follows:

$$\epsilon ={\displaystyle \frac{1}{n}}\sum _{i=1}^n\left|{\displaystyle \frac{Y_i-Y_i^{reg}}{Y_i}}\right|,$$

(3)

where n—sample size, Yᵢ—observed response function values obtained from the experimental and literature data, Y_i^reg—response function values calculated using the regression equation, i—sample number.

The coefficient of determination (R²), which indicates the significance of the dependence of the response functions on the process factors and ranges from 0 to 1, was determined according to the methods described in [39].

3. Results and Discussion

3.1. Database Creation

As a fundamental assumption in developing the equations, the authors proposed that the yield of humic acids primarily depends on the chemical composition of brown coal, which, in turn, is determined by the degree of lignite coalification and the depositional environment. Therefore, to determine the HA yield based on the technical characteristics of lignite, the development of the regression empirical equation (experimental–statistical model, ESM) employed volatile matter yield (V) and higher heating value (Q). The amount of volatile matter and heating value depends on the elemental composition of coal and the combination of these elements, i.e., it characterizes the chemical composition of lignite. Considering that only a small amount of ash is transferred from lignite to HA [14], it was decided to use these parameters relative to the ash-free sample (V^af, wt.%; Q^af, kJ/kg). Additionally, HA contains moisture, which is influenced by the content of functional groups in HA and the moisture content in coal (more than 50% of the moisture in coal can be concentrated in HA). Therefore, the moisture content in lignite on an analytical sample basis (W^a, wt.%) was also used as a characteristic of coal that affects the yield of HA.

As noted in Section 2, the development of the ESM for the dependence of humic acid yield on coal characteristics was based on data from Ukrainian lignite samples obtained in the authors’ previous studies [27] and from non-Ukrainian deposits presented in [28,29,30,31]. Based on these source data, the main physicochemical characteristics of lignite and HA content were calculated and compiled, as shown in

Table 5. Input data for developing the mathematical model for predicting HA yield.

№	Wa, wt.%	Vaf, wt.%	Qaf, kJ/kg	HAaf, wt.%
	X1i	X2i	X3i	Yi
1	16.50	41.89	22,156	63.50
2	8.70	45.38	20,062	44.91
3	30.40	41.29	12,708	49.12
4	4.32	45.38	12,971	58.33
5	2.61	40.43	12,329	64.06
6	5.33	44.59	12,192	57.36
7	3.44	44.29	13,679	61.01
8	3.78	42.76	13,128	62.61
9	4.44	35.68	12,763	53.68
10	9.63	49.00	17,740	43.04
11	7.70	47.54	17,124	44.95
12	23.00	39.53	12,289	40.70
13	9.30	37.95	15,607	51.86

.

3.2. Development of the Regression Empirical Equation

To select the type of equations, Figure 1 presents the nature of the influence (correlation fields) of coal characteristics (X) on the response function (Y). Based on the appearance of the correlation fields (Figure 1), a linear regression pattern was assumed. Therefore, a polynomial of the form Y = a₀ + a₁∙X₁ + a₂∙X₂ +a₃∙X₃ was used as the mathematical model (notations Y, X₁, X₂, X₃ are provided in Table 5), with the unknown coefficients a₀, a₁, a₂, and a₃ determined by the least squares method.

The following equation was obtained:

Y = 85.6718643370032 − 0.498874171045942∙X₁ − 0.627768036640667∙X₂ − 0.0000273053461338392∙X₃

(4)

3.3. Validation of the ESM Adequacy

For each sample, the predicted response function values (Y_i^reg) were calculated by substituting the values of X₁–X₃ into the above equation, and the relative errors of the ESM (ε_i) were determined, as presented in

Table 6. Experimental data, calculated response function values, and relative errors.

Sample No.	X1, wt.%	X2, wt.%	X3, kJ/kg	Y, wt.%	Yireg, wt.%	εi
1.	16.50	41.89	22156	63.50	50.54	0.2565
2.	8.70	45.38	20062	44.91	52.30	0.1412
3.	30.40	41.29	12708	49.12	44.24	0.1103
4.	4.32	45.38	12971	58.33	54.67	0.0669
5.	2.61	40.43	12329	64.06	58.65	0.0922
6.	5.33	44.59	12192	57.36	54.69	0.0489
7.	3.44	44.29	13679	61.01	55.78	0.0938
8.	3.78	42.76	13128	62.61	56.58	0.1065
9.	4.44	35.68	12763	53.68	60.71	0.1158
10.	9.63	49.00	17740	43.04	49.62	0.1327
11.	7.70	47.54	17124	44.95	51.52	0.1275
12.	23.00	39.53	12289	40.70	49.05	0.1702
13.	9.30	37.95	15607	51.86	56.78	0.0867
Average relative approximation errors (ε)						0.1192

.

The adequacy of the developed model was verified using the methodologies described in Section 3.1, based on the regression response functions. The following patterns were identified when validating the equations (Equation (4)). The majority of the residuals ∆Y_i = Y_i^reg – Y_i, are illustrated in the histograms and probit plots (Figure 2 and Figure 3).

The histogram of residual distribution overlaid with a standard distribution curve and the probit plot indicates minor deviations of the residual distribution from the normal distribution law.

The average relative approximation error is ε = 0.1192 (11.92%). According to the recommendations in [40], when ε = 0–10%, the prediction accuracy is considered high, 10–20% is good, and 20–50% is satisfactory. Based on this, it can be concluded that the developed models exhibit good agreement with the experimental data.

The coefficient of determination is R²₁ = 0.29064. This indicates that 29% of the variation in the yield of humic acids is determined by the coal characteristics, which are taken as influencing factors, possibly due to the relatively limited amount of experimental data used to develop the ESM. But the correlation coefficient R = 0.53911 confirms the assumption of a significant linear relationship (in 54 cases out of 100) between the coal characteristics, which are taken as influencing factors on the yield of humic acids, and the yield of humic acids, which can be considered a satisfactory result [41].

Given that the difference between the calculated and experimental HA yields is relatively small, but the relationship between coal characteristics and the response function is questionable, a reverse verification of Equation (4) was conducted to determine the feasibility of using this equation for forecasting HA yield. To do this, the theoretical HA yield was calculated based on the characteristics of lignite from the Morozivske deposit, which were not used in developing the ESM, and compared to the experimental results (see

Table 7. Calculated HA content in the Morozivske deposit.

Moisture Content, Wa, wt.%	Volatile Matter Content, Vaf, wt.%	Higher Heating Value, Qaf, kJ/kg	Calculated Content HA, HAreg, af, wt.%	HA Content, HAaf, wt.%	Relative Approximation Error, %
X1	X2	X3	Yreg	Y	ε
26.40	43.80	16458	44.56	48.33	7.80

).

The presented data show that the predicted HA yield is slightly lower than the experimental value. Still, the prediction error is within 8%, which can be considered a reasonably good result.

3.4. Characteristics and Humic Acid Yields of Ukrainian Lignite

The obtained equations were used to calculate the humic acid content. Reference data [33] on the characteristics of Ukrainian lignite were used as input data for the calculations. The overall HA yield from lignite extracted in various regions of Ukraine is presented in

Table 8. Calculated HA content in Ukrainian lignite.

No.	Region Name	HA Content
		HAaf, wt.%	HAa, wt.%	HAdaf, wt.%
1	Dnipropetrovsk Region (Dnieper-Donets Basin, Southern Flank)	52.77–55.83	38.52–49.46	59.13–62.72
2	Dnipropetrovsk Region (Ukrainian Shield, Central Uplift, Prydniprovsky Block)	44.03–45.92	35.44–40.44	51.18–74.57
3	Zhytomyr Region (Ukrainian Shield, Central Uplift, Volyn Block)	47.34–47.42	36.40–37.30	56.19–62.01
4	Zakarpattia Region (Transcarpathian Intermountain Basin, Chop–Mukachevo Zone)	45.91–49.19	35.77–37.87	52.87–62.03
5	Kirovohrad Region (Ukrainian Shield, Central Uplift, Kirovohrad Block)	40.86–48.48	31.54–40.05	47.36–62.72
6	Kharkiv Region (Dnieper-Donets Coal-Bearing Area)	44.94–47.14	32.53–41.05	52.57–56.72
7	Cherkasy Region (Vatutine Geological and Industrial District)	43.17–51.59	31.15–42.95	46.11–58.63

. As seen from Table 8, the highest predicted humic acid content, in terms of potential content in mined coal (HA^a) or enriched and dried coal (HA^daf), is observed in the lignite of Dnipropetrovsk Region. Zhytomyr and Zakarpattia Regions also note high values of HA^a and HA^daf. Cherkasy Region shows moderate levels of humic acids, with slightly lower HA^a and HA^daf values. The lowest HA^a values are observed in Kirovohrad and Kharkiv Regions, indicating lower efficiency for humic acid extraction from lignite in these areas. However, certain local deposits may still possess significant industrial potential.

The physicochemical characteristics of HA from selected Ukrainian lignite samples are presented in

Table 9. Physicochemical characteristics of humic acids.

HA Name	Wa, wt.%	Ad, wt.%	Vdaf, wt.%	Elemental Composition, wt.%
				Cdaf	Hdaf	Ndaf	Sdaf	Odaf
From Sample No. 1 of the Mokrokalyhirske lignite open-pit mine	8.50	13.50	43.20	63.13	3.90	1.46	4.20	27.31
From Sample No. 2 of the Mokrokalyhirske lignite open-pit mine	5.30	8.20	41.00	58.20	3.70	3.80	3.13	31.17
From Sample No. 3 of the Mokrokalyhirske lignite open-pit mine	12.10	11.00	55.80	63.81	3.70	1.23	0.76	30.50
From the Morozivske deposit sample	8.00	4.30	50.30	62.30	3.86	0.79	4.79	28.26

.

The results in Table 9 indicate that, compared to the characteristics of the original coal (see Table 2 and Table 3), the HA are characterized by the following:

-

Lower moisture and ash content;

-

Lower volatile matter yield;

-

Higher oxygen and sulfur content.

Compared to the raw lignite, this suggests a lower ballast content, greater thermal stability, and a higher concentration of heteroatomic compounds in the humic acids, which translates to improved surface-active properties.

4. Conclusions

During the study, an experimental–statistical model was successfully developed to forecast the yield of humic acids (HA) from lignite based on moisture, volatile matter yield, and calorific value. The model’s adequacy verification showed an average relative approximation error of approximately 12%, indicating sufficiently high prediction accuracy. The most promising for humic acid production is the Dnipropetrovsk Region (HA yield is 35–50 wt.% on an analytical sample basis and 51–75 wt.% on the organic part basis). The results highlight the significant potential of lignite as a source of humic substances, which are widely applicable in agriculture, environmental protection, and various technological fields. HA are characterized by low moisture and ash content, and a high content of heteroatomic functional groups, making them suitable as high-quality surface-active, antibacterial, and antioxidant additives for biodegradable polymers, bitumen, sorbents, and more.

However, the study also indicated a limited correlation between the selected parameters and HA content, emphasizing the need to consider additional factors to enhance prediction accuracy. The developed model is suitable for preliminary evaluation of HA content in Ukrainian lignite, but it should be validated with local data from other geological regions. The authors hope that the study data will encourage the resumption of lignite mining in Ukraine, which will allow more samples to be collected to improve the established dependence of the yield of humic acids on the characteristics of brown coal.