Closed-Loop Chemical Recycling of Polylactide via Glycolysis: From Water-Soluble Oligomers to High-Purity Lactide

Abstract

Polylactide (PLA) has become widely adopted across biomedical, packaging, and manufacturing sectors due to its biodegradability and renewable sourcing. However, the rapid growth in PLA consumption has created urgent challenges related to waste management and the cleaning of processing equipment. This study investigates glycolysis as a promising chemical depolymerization pathway for PLA recycling and in situ reactor cleaning. A systematic analysis of four glycolysis agents (GA) (ethylene glycol, diethylene glycol, propylene glycol, and glycerol) was performed across molar PLA:GA ratios from 1:0.125 to 1:4 at 220 °C, targeting the efficient conversion of high-molecular-weight PLA (Mn ≈ 165 kDa) into low-molecular-weight oligomers. Gel permeation chromatography (GPC) demonstrated that propylene glycol exhibited the highest depolymerization efficiency, yielding oligomers with Mn as low as 200 g·mol⁻¹ even at minimal glycolysis agent ratios, while glycerol produced hydroxyl-rich oligomers optimal for subsequent lactide synthesis. Hydroxyl value (HV) measurements showed excellent agreement with theoretical values (<5% deviation), allowing us to make an assumption about an approximate, close to near-quantitative con-version. Glycolysis products with Mw below 400 g·mol⁻¹ displayed excellent water solubility, making them particularly attractive for reactor cleaning applications. Using glycerol-derived (GL) oligomers (PLA:GL = 1:0.25), purified L-lactide with a melting point of 98.1 °C and high purity (>99%) was obtained through thermocatalytic depolymerization and five recrystallization cycles, as confirmed by ¹H nuclear magnetic resonance (¹H NMR) and differential scanning calorimetry (DSC) analyses. The recovered lactide’s high purity renders it suitable for ring-opening polymerization, enabling closed-loop PLA recycling schemes. Overall, glycolysis emerges as a highly promising chemical recycling route complementary to hydrolysis and pyrolysis: propylene glycol maximizes depolymerization efficiency for cleaning applications, while glycerol optimizes oligomer functionality for lactide recovery and advanced material synthesis. Our results provide practical guidelines for selecting glycolysis agents and conditions for cleaning and recycling applications.

1. Introduction

In recent years, there has been an increasing interest in biodegradable polymers, both in academic and industrial circles. Their derivation from renewable resources and their ability to degrade with minimal environmental impact are significant advantages over oil-based polymers. Biodegradable polymers are used in various aspects of human life, including agriculture, packaging, and medicine [1,2].

The growing interest, increase in production, and tighter environmental regulations are all leading to an increased use of products made from biodegradable polymers. However, this rising popularity has raised questions about how such products should be handled at end-of-life [3,4,5]. Although biodegradability implies spontaneous degradation under certain environmental conditions, these conditions are not always met. As a result, when polylactide products are disposed of in urban landfills or aquatic environments, the conditions required for rapid decomposition are not satisfied, leading to the accumulation of waste and further pollution [6,7,8,9,10,11]. Thus, there is a clear need for proper disposal and recycling methods for these products.

Polylactide (PLA) is a thermoplastic aliphatic polyester and one of the most sought-after biopolymers today [12]. Its most promising application is the creation of bioabsorbable medical implants, which can be used in various medical fields, including orthopedics, dentistry, maxillofacial surgery, and general surgery [13]. In addition, PLA has found widespread use in single-use food packaging, automotive manufacturing, mechanical engineering, and 3D printing [14].

The main method of synthesizing high-molecular-weight polylactide is ring-opening polymerization of the lactide monomer in batch reactors [15,16,17]. This raises the issue of reactor purification from residual high-molecular-weight polylactide after synthesis. The insolubility of PLA in water severely restricts the choice of solvents for reactor cleaning, often necessitating the use of toxic chlorinated solvents such as chloroform, dichloromethane, and dichloroethane [18].

The accumulation and disposal of polylactic acid waste, as well as the cleaning of processing equipment, are further complicated by the relatively high cost of lactic acid obtained through fermentation and purification. These factors make PLA processing and reuse an important technological and economic issue [3].

Several methods have been proposed for the utilization of polylactide, including industrial composting, mechanical recycling, and chemical recycling [3]. Composting under industrial conditions leads to rapid degradation, but the products of this process are primarily carbon dioxide and water, which does not fit the concept of feedstock reuse [19,20,21,22]. Mechanical recycling generally leads to a lower-molecular-weight polymer with degraded properties due to chain scission [23,24,25]. Chemical recycling appears to be the most promising approach, owing to its ability to regenerate the initial chemical building blocks that can be reused in PLA production.

The literature describes various approaches to the chemical recycling of polylactide. Three main methods are commonly distinguished [3]: hydrolysis in water to produce lactic acid [26,27,28,29,30]; alcoholysis to yield alkyl lactates [31,32,33]; and pyrolysis of polylactic acid [34,35,36,37,38]. However, relatively little attention has been paid to glycolysis, despite its potential as a versatile chemical recycling method [39,40,41,42,43]. Scheme 1 shows mechanisms of chemical processing of polylactide by glycolysis, hydrolysis, alcohololysis and pyrolysis.

Previous studies have explored PLA glycolysis under specific conditions. For instance, Jaikaew [44] investigated microwave-assisted glycolysis with Fe₃O₄ nanoparticle catalysts, focusing on process optimization for a single glycolysis agent—ethylene glycol, but without systematic comparison of multiple glycols. Similarly, Paciorek-Sadowska et al. [44] utilized glycolysis with diethylene glycol to generate low-molecular-weight PLA for polyurethane-isocyanurate foams, demonstrating effective depolymerization but limiting the scope to one specific glycols. Nim et al. [45] conducted microwave-assisted alcoholysis of PLA with various diols, achieving quantitative product analysis and identifying optimal conditions for oligomer formation; however, their work emphasized microwave heating, different catalysts, and alkyl lactate endpoints rather than conventional thermal glycolysis. However, no systematic study has been conducted on this process, its structure, and properties, leaving a gap in this area.

Glycolysis is the process of molecular depolymerization through transesterification of PLA ester groups. During glycolysis, ester bonds are cleaved by nucleophilic attack of glycol molecules [42]. While hydrolysis produces lactic acid and alcoholysis yields alkyl esters with established recycling pathways, glycolysis offers a distinct advantage: direct conversion to lactide-compatible oligomers without intermediate purification steps. Low-molecular-weight polyols and other liquid products generated during glycolysis can be used to produce lactide, which can then be repolymerized to polylactide, forming a closed loop.

This characteristic makes glycolysis uniquely suited for both equipment decontamination and feedstock recovery. Furthermore, glycolysis operates at lower temperatures (220 °C vs. 240–280 °C for pyrolysis) and generates water-miscible products, reducing solvent waste and environmental burden [46].

Based on these considerations, there is a clear need for environmentally friendly and resource-efficient methods for PLA recycling. Glycolysis, as a form of chemical depolymerization, offers a promising route to reduce PLA molecular weight to reactive oligomers that can be further transformed into lactide and regenerated PLA, enabling a closed-loop recycling scheme in which glycolysis-derived oligomers serve directly as feedstock for lactide synthesis and subsequent ring-opening polymerization.

This work systematically evaluates comparative process guidelines of PLA glycolysis using four structurally diverse polyols in order to establish practical guidelines for controlling oligomer molecular weight distribution and water solubility, as well as for selecting oligomer structures that are optimal for downstream lactide production. We hypothesize that polyol structure (diol vs. triol, chain length and polarity) significantly influences reaction kinetics and product characteristics, with direct implications for industrial-scale PLA recovery, equipment cleaning, and efficient conversion of glycolysis products into high-purity lactide.

2. Materials and Methods

2.1. Materials

Polylactide Ingeo biopolymer 4032D (PLA, NatureWorks, Plymouth, MN, USA) was used as the starting polymer. L-polylactide with molecular weight about 165 kDa and melting temperature about 160 °C.

Ethylene glycol (EG, 99.5%, KomponentReactiv, Moscow, Russia), diethylene glycol (DEG, 99.5%, KomponentReaktiv, Moscow, Russia), propylene glycol (PG, 99.5%, KomponentReaktiv, Moscow, Russia) and glycerol (GL, 99.3%, KomponentReaktiv, Moscow, Russia) were used as glycolysis agents. Zinc Stearate (ZnSt, zinc content not less than 10.5%, BELIKE Chemical Co., Ltd., Zhongshan, China) was used as a catalyst for the glycolysis process. Butyl acetate and chloroform (>99%) were purchased from Ekos-1 (Staraya Kupavna, Russia) and were used without further purification. The catalyst tin octoate (Dabco T-9; Evonik Industries AG, Staufen, Germany) and 1,12-dodecanediol (Suzhou Senfeida Chemical Co., Ltd., Changzhou, China) were used without further purification.

2.2. Glycolysis Process

The glycolysis reaction was carried out in a three-necked flask equipped with an overhead stirring device with an anchor-type agitator. The process temperature was set at 220 °C in an argon atmosphere, and heating was carried out using a flask heater (Joanlab 22403-HMC, JOAN LAB Equipment Co., Ltd., Huzhou, China) with an accuracy of ±1 °C.

Based on our previous experience in the process of polymer glycolysis [40,41] and the literature data on the thermal behavior of polylactide [47], we selected these conditions.

Temperature control was achieved through the use of a thermocouple, and the reaction time ranged from 60 min to 180 min depending on the ratio of glycolysis agent to initial polymer. The system was stirred at 50 rpm throughout the process.

First, polylactide was melted in the flask at the reaction temperature. Then, the catalyst and the glycolysis agent, preheated to 220 °C, were introduced into the reaction space. The starting point of the reaction was taken as the time of glycolysis agent injection.

The calculated amount of glycolysis agent, polymer, and catalyst were added to the flask. The reaction mixture was purged with an inert gas for at least 15 min and then placed in the flask heater for heating. The formulations of the reaction mixtures are given in

Table 1. The formulations of reaction mixtures.

№	Component	Name	Mass, g
1	Polymer	PLA 4032D	20
2	Glycolysis agent	EG, DEG, PG, GL	Z*
3	Catalyst	Zinc stearate	0.2

Z*—the amount of glycolysis agent calculated by Equation (1).

.

The initial mass of each glycolysis agent was calculated according to Equation (1):

$$\mathrm{m}\left(1\right)={\displaystyle \frac{\mathrm{m}\left(2\right)}{M\left(2\right)}}·M\left(1\right)·\mathrm{z},$$

(1)

where m(1)—mass of the glycolysis agent, m(2)—mass of the polylactide, M(1)—molecular weight of the glycolysis agent, M(2)—molecular weight of the polylactide, and z—specified PLA:GA ratio.

The optimal temperature for comparing the effectiveness of all glycolysis agents was chosen as 220 °C. At this temperature, a homogeneous (single-phase) process occurs. PLA 4032D begins to melt at a temperature of 160 °C, however, in the temperature range from 160 to 210 °C, melting proceeds slowly and a fully homogeneous phase is not observed. At process temperatures from 240 to 260 °C, significant polylactide degradation was observed during prolonged glycolysis. These data converge with optimal processing intervals based on data on the thermal behavior of the polymer indicated in the literature [47]. Most glycolysis agents, except glycerol and diethylene glycol, have boiling points below 230 °C, which complicates operation in this higher temperature range. The choice of glycolysis agents was based on their availability and ability to mix with water.

Zinc stearate was chosen as the catalyst because zinc carboxylates are well-known Lewis acid catalysts for esterification and transesterification reactions. In addition, zinc stearate is inexpensive, commercially available, thermally stable under our reaction conditions, and easy to handle.

The optimal time for glycolysis of PLA is 60 min for the ratios 1:0.5; 1:1; 1:2; 1:4. For the ratios below, the process time has been increased to 120 and 180 min, in some cases.

Figure 1 shows a laboratory setup for glycolysis process of polylactide.

2.3. Synthesis of Lactide

Lactide synthesis was carried out on the same setup and using the same technology as in our previous work [2]. Polycondensation of lactic acid was performed in a 500 mL three-necked flask equipped with a magnetic stirrer and connected to a receiving flask fitted with a straight condenser. The product obtained after glycolysis was cooled, filtered from mechanical impurities, and introduced into the flask to produce lactide via thermocatalytic depolymerization of lactic acid oligomers.

The flask was purged with nitrogen (99.96% purity) during the process, and the temperature was maintained between 140 °C and 200 °C using a flask heater (Joanlab 22403-HMC, JOAN LAB Equipment Co., Ltd., Huzhou, China) with an accuracy of ±1 °C.

The polycondensation process was carried out at a gradually decreasing pressure from 510 down to 10 mbar over three hours. Samples were collected at certain time points to analyze the molecular weight. The mass and refractive index of the vaporized water were used to determine the conversion stage. Once the release of water from the reaction flask ceased, the receiving flask was changed, and the temperature was raised to 210 °C. The pressure in the reaction system remained between 5 and 10 mbar for an hour and a half. The raw lactide was then isolated under vacuum and transferred to a receiving flask placed in a water bath maintained at 10 °C.

Figure 2 shows a laboratory setup for the production of lactide through thermocatalytic depolymerization of lactic acid oligomers.

The purification of the obtained raw lactide was accomplished through the process of recrystallization using butyl acetate as the solvent. The procedure was conducted in a 100 mL beaker placed on a heating plate equipped with a magnetic stirrer. The ratio of lactide to solvent was 3:1. The temperature was maintained at 70 °C, and the stirring was provided by the heating plate, IKA C-MAG HS-7 (IKA-Werke GmbH & Co., KG, Staufen, Germany). Once the substance had fully dissolved, the beaker was removed from the hotplate and allowed to cool to 5 °C. The lactide crystals were then separated from the solvent using Buchner funnel vacuum filtration. The process was repeated 5 times.

2.4. Differential Scanning Calorimetry (DSC)

The thermal effect of lactide was studied with a differential scanning calorimeter, Netzsch DSC 204 (NETZSCH GmbH & Co. Holding KG, Selb, Germany), in a dry inert atmosphere (argon 99.99%) at a flow rate of 20 mL/min in the range from 10 to 260 °C with a heating rate of 10 °C/min. A prepared sample weighing from 4 to 7 mg was taken and placed in a standard aluminum crucible with a lid. The results were processed using the Netzsch Proteus program (version 5.2.1). The software was used to determine the melting onset temperature, extremum temperature and thermal effect (area under the peak).

2.5. Vibration Viscometry (VV)

The viscosity of glycolysis products was determined by vibrational viscometry using a viscosity analyzer AND SV-10 (A&D Co., Ltd., Tokyo, Japan). The measurement method is based on tuning fork vibration at a frequency of 30 Hz. Measurements were performed at 25 °C. The minimum sample volume is 20 mL, and the measuring range from 0.3 to 10,000 MPa·s. The samples were poured into a special glass cuvette.

2.6. Gel Permeation Chromatography (GPC)

Molecular weight characteristics were investigated by gel permeation chromatography (Gilson Inc., Villiers-le-Bel, France). The analysis was carried out at 25 °C in tetrahydrofuran (THF) at a flow rate of 1.0 mL/min with a refractive index detector. For the separation and identification of oligomeric fractions, a PLgel 3 µm MIXED E column (separating capacity of up to 30 kDa) (Agilent, Santa Clara, CA, USA) was used, with a set of polystyrene standards with Mp (peak molecular weight) of 580, 1280, 2940, 10,110 and 28,770 g/mol and PDI lower than 1.12 (Agilent). For the separation and identification of high-molecular-weight polymers, a PLgel 5 µm MIXED B column (separating capacity of 500–3,000,000 Da) (Agilent, Santa Clara, CA, USA) was used, with a set of polystyrene standards with Mp (peak molecular weight) of 2940, 10,110, 28,770, 74,800, 230,900 and 1,390,000 g/mol and PDI lower than 1.12 (Agilent). A solution of polymer in eluent was prepared for analysis with polymer concentration not exceeding 5 mg/mL and not less than 1 mg/mL.

2.7. Nuclear Magnetic Resonance (NMR)

The qualitative and quantitative composition of lactide and PLA was studied by using an NMR spectrometer (¹H NMR), BrukerDPX-500 (Bruker Corporation, Bremen, Germany), with Fourier transform in CDCl₃ (internal standard: TMS).

2.8. Hydroxyl Value (HV)

The hydroxyl value (HV) of the obtained oligomers was determined in accordance with DIN 53240, which is used for polyols, polyesters, alkyd resins, and other compounds containing hydroxyl groups (-OH). The hydroxyl value was determined by acetylation, i.e., reaction of hydroxyl groups with acetic anhydride followed by titration of excess anhydride with alkali. In this work, propionic anhydride was used instead of acetic anhydride, and the reaction time was increased to ensure complete conversion. The method was calibrated against standards with known HN values, with an error of ±5 mg KOH/g.

The theoretical hydroxyl number was calculated using Equation (2):

$${\mathrm{H}\mathrm{V}}_{theor}={\displaystyle \frac{\mathrm{56,100}·\mathrm{f}}{M_n}},$$

(2)

where 56,100—coefficient related to the molecular weight of KOH (56.1 g/mol × 1000 mg/g), f—functionality (number of -OH groups per molecule), M_n—number average molecular weight.

3. Results and Discussion

3.1. Polylactide Glycolysis

During glycolysis, the preheated glycolysis agent was added to the molten PLA. This procedure was chosen because heating all components together leads to inhomogeneous reaction that begins at the polymer surface as it softens.

Table 2. Conditions of the glycolysis process and molecular weight distribution of the obtained oligomers.

№	Glycolysis Agent	PLA:GA Ratio	Reaction Time, min	Mn, Da	Mw, Da
1	Ethylene glycol	1:0.25	180	550	690
2		1:0.5	120	230	280
3		1:1	60	170	190
4		1:2	60	130	150
5		1:4	60	110	130
6	Diethylene glycol	1:0.125	120	750	1980
7		1:0.25	60	710	840
8		1:0.5	60	290	340
9		1:1	60	200	230
10		1:2	60	150	180
11		1:4	60	200	230
12	Propylene glycol	1:0.125	120	1040	3140
13		1:0.25	60	650	1150
14		1:0.5	60	350	440
15		1:1	60	230	290
16		1:2	60	220	270
17		1:4	60	200	240
18	Glycerol	1:0.25	120	460	600
19		1:0.5	60	250	310
20		1:1	60	170	200
21		1:2	60	130	170
22		1:4	60	150	160

shows the glycolysis conditions and the molecular weight distributions of the resulting oligomers.

At PLA:GA ratios lower than 1:1, the reaction mixture remained homogeneous and appeared as a molten PLA phase containing the glycolysis agent. When the ratio exceeded 1:1, phase separation occurred, forming a suspension of PLA melt droplets in the glycolysis agent.

In all cases, the mass of the reaction system remained almost unchanged during glycolysis. According to gel-penetrating chromatography, the glycolysis process began almost immediately after the introduction of the glycolysis agent.

As summarized in Table 2, the evolution of Mn and Mw with increasing PLA:GA ratio does not always follow a strictly monotonic trend, which can be rationalized by the interplay of several concurrent processes. At low glycolysis-agent loadings, incomplete transesterification and limited phase homogeneity lead to broader molecular-weight distributions and apparently higher Mn values, whereas at intermediate ratios efficient chain scission dominates and narrow oligomer populations are obtained. At the highest glycolysis-agent excess, dilution by unreacted polyol and partial re-aggregation or secondary transesterification between short chains can slightly increase Mw or broaden dispersity, giving rise to the apparent “plateau” and minor deviations from strictly decreasing trends seen for some systems in Table 2.

Figure 3 shows a graph of the dependence of the drop in the average molecular weight on the ratio of the glycolysis agent.

As shown in Figure 3, using glycolysis agents at ratios above 1:1 relative to PLA does not lead to a further significant decrease in the molecular weight of the products, and the dependence of the average molecular weight on the glycolysis agent ratio reaches a plateau.

When ethylene glycol was used as the glycolysis agent, a significant decrease in molecular weight was observed at a PLA:GA ratio of 1:1. However, further increases in the amount of glycolysis agent did not produce a substantial additional decrease in molecular weight. At this ratio, the transition from high molecular weight to low molecular weight products occurred within 15–30 min. The molecular weight then changed only slightly with further reaction time, as shown by the GPC chromatograms in Figure 4.

Based on the GPC curves in Figure 5 for products obtained at different PLA:EG ratios, it can be concluded that at PLA:GA = 1:0.25 a reaction time of 180 min is not sufficient to obtain low-molecular-weight products. At ratios of 1:2 and 1:4, two peaks are observed in the chromatograms, one of which corresponds to pure EG due to excess glycolysis agent.

When DEG was used as the glycolysis agent, conversion from high-molecular-weight to low-molecular-weight product did not occur within the selected process times at PLA:GA ratios of 1:0.125 and 1:0.25. This can be seen in the GPC curves in Figure 6. However, at an excess of glycolysis agent (PLA:DEG = 1:4), a larger fraction of low-molecular-weight products was observed.

During glycolysis with propylene glycol, low-molecular-weight products were formed even at a PLA:PG ratio of 1:0.125, as shown by the GPC curves in Figure 7. However, the average molecular weight remained relatively high due to the presence of broad oligomer distributions. At PLA:PG ratios of 1:1, 1:2, and 1:4, the peak corresponding to pure propylene glycol increased in intensity, and the average molecular weight decreased with increasing glycolysis agent ratio.

According to the GPC data in Figure 8, using glycerol as the glycolysis agent at PLA:GL = 1:0.125 was insufficient to produce low-molecular-weight products. At ratios of 1:2 and 1:4, a peak corresponding to pure glycerol appeared due to the excess glycolysis agent.

Figure 9 shows the time evolution of GPC curves for glycolysis with glycerol at a PLA:GL ratio of 1:2.

Chromatograms were recorded after 10, 15, and 30 min (oligomeric column) and 60 min (high-molecular-weight column). Analysis of the curves shows that the peak of pure glycerol (~90 Da) decreases with increasing glycolysis time on the oligomeric column, while the peak of a low-molecular-weight product (~300 Da) increases. On the high-molecular-weight column, the resulting product is represented by two peaks, one corresponding to pure glycerol and the other to a product with a molecular weight of approximately 220 Da. This product corresponds to one PLA repeat unit and one glycerol molecule.

As can be seen from the GPC curves for the different glycolysis agents, macromolecules are broken down into smaller molecules, followed by the formation of n-mers (oligomers with different degrees of polymerization). This is reflected in the characteristic “jagged” appearance of the curves.

The comparative glycolysis kinetics reveal a clear reactivity hierarchy: propylene glycol > ethylene glycol > glycerol > diethylene glycol. At stoichiometric ratios (PLA:GA = 1:1), propylene glycol achieved Mn ≈ 230 Da within 60 min, whereas diethylene glycol required substantially longer reaction times at comparable PLA:GA ratios to reach similarly low molecular weights (Table 2). This differential reactivity is associated with differences in polyol structure and miscibility. The higher apparent reactivity of propylene glycol compared to ethylene and diethylene glycols may arise from a combination of steric and polarity effects and, most importantly, from its superior miscibility with the PLA melt. This enhanced miscibility maintains a higher local concentration of reactive hydroxyl groups at the ester sites, thereby facilitating transesterification. Conversely, glycerol’s three hydroxyl groups compete for single ester sites, slowing per-hydroxyl conversion kinetics despite its apparent reactivity at low PLA:GA ratios. The apparent anomaly at PLA:DEG = 1:0.125 (Mn = 750 Da after 120 min) suggests incomplete reaction competing with oligomerization via polycondensation, consistent with literature reports of side reactions at low glycolysis agent concentrations.

The overall reactivity order (PG > EG > GL > DEG) can thus be rationalized by differences in nucleophilic attack rates and phase behavior. Detailed kinetic analysis, however, lies beyond the scope of this study and would require dedicated investigations. In the homogeneous phase region (PLA:GA < 1:1), PG remains fully miscible, maintaining a high local concentration of reactive nucleophiles at the ester sites. In contrast, DEG and EG tend to form biphasic systems at higher glycolysis agent ratios, reducing interfacial contact and slowing reaction kinetics.

3.2. Dynamic Viscosity

In the context of using the glycolysis method for cleaning technological equipment, as well as for characterizing glycolysis products, the dynamic viscosity of the final system was measured.

Table 3. Results of viscosity measurements of glycolysis products.

№	Glycolysis Agent	PLA:GA Ratio	Dynamic Viscosity, Pa·s	Mixing with Water
1	Ethylene glycol	1:0.25	>10	Bad
2		1:0.5	-	Good
3		1:1	0.19	Good
4		1:2	0.11	Good
5		1:4	0.11	Good
6	Diethylene glycol	1:0.125	>10	Bad
7		1:0.25	3.59	Bad
8		1:0.5	0.37	Good
9		1:1	0.10	Good
10		1:2	0.06	Good
11		1:4	0.07	Good
12	Propylene glycol	1:0.125	10.5	Bad
13		1:0.25	3.23	Bad
14		1:0.5	0.80	Good
15		1:1	0.60	Good
16		1:2	0.10	Good
17		1:4	0.80	Excellent
18	Glycerol	1:0.25	>10	Good
19		1:0.5	>10	Good
20		1:1	>10	Good
21		1:2	3.77	Good
22		1:4	1.27	Excellent

shows the results of viscosity measurements of the end products of glycolysis.

In order for the glycolysis process to be effective under industrial conditions, it is necessary to select a glycolysis agent whose degradation products are soluble in water and do not remain on reactor walls.

In this study, 10 wt.% aqueous solutions of glycolysis products were prepared and stirred at 90 °C. The resulting solution was then transferred to a flask, and its appearance was recorded immediately and after three days. A rating of “Excellent” corresponded to complete mixing of glycolysis products and water; “Good” indicated dispersion with varying degrees of flake formation; and “Bad” corresponded to complete separation of the glycolysis products and water.

Low viscosities observed for samples obtained with excess glycolysis agent can be explained by the dilution effect of unreacted glycolysis agent.

Figure 10 shows the results of a test for mixing glycolysis products with water.

When glycolysis products were mixed with water at PLA:GA ratios of 1:0.5 or higher, mixing occurred in all cases, regardless of the glycolysis agent used. However, turbidity of the solutions was often observed. This can be attributed to the limited solubility of lactic acid oligomers in water, which decreases as oligomer molecular weight increases. Therefore, in all cases where the molecular weight was below 400 g·mol⁻¹, the glycolysis products exhibited relatively good solubility. This is consistent with literature data on the solubility of lactic acid oligomers in water [48].

The presence of long-chain molecular remnants, mechanical impurities such as unreacted PLA particles resulting from technological imperfections, and residues of unreacted glycolysis agent all contribute to the turbidity of the final solutions. These residues may adhere to the walls of the reactor or agitator above the level of the reaction mixture. After three days, in most cases, the turbidity disappeared due to additional hydrolysis of oligomer molecules by water, which further decreased their molecular weight. In some cases, stratification and precipitation of PLA particles were observed. Such phase separation can be explained by the limited solubility of excess glycolysis agents in water.

The correlation between oligomer molecular weight and aqueous solubility reflects the established solubility behavior of lactic acid oligomers: oligomers with molecular weights below about 400 Da are hygroscopic due to unmasked hydroxyl functionality, whereas those with molecular weights above about 1000 Da become increasingly hydrophobic.

3.3. Hydroxyl Value

To evaluate the extent of the glycolysis reaction and the degree of chain scission, the hydroxyl values of the obtained oligomers were determined according to Section 2.8.

Table 4. Hydroxyl values of glycolysis products and comparison with theoretical values.

№	Glycolysis Agent	PLA:GA Ratio	Hydroxyl Value, mgKOH/g	Theoretical Hydroxyl Value, mgKOH/g	Mn, Da
1	Ethylene glycol	1:0.25	317	320	550
2		1:0.5	530	544	230
3		1:1	821	836	170
4		1:2	1142	1144	130
5	Diethylene glycol	1:0.25	301	285	710
6		1:0.5	391	448	290
7		1:1	583	630	200
8		1:2	754	789	150
9	Propylene glycol	1:0.25	287	308	650
10		1:0.5	549	510	350
11		1:1	908	758	230
12		1:2	1038	1001	220
13	Glycerol	1:0.25	461	443	460
14		1:0.5	655	712	250
15		1:1	972	1025	170
16		1:2	1254	1313	130

presents the experimental hydroxyl values, theoretical hydroxyl values, and corresponding Mn values for selected systems. Figure 11 shows the comparison between experimental and theoretical hydroxyl values.

Hydroxyl value (HV) measurements constitute the primary quantitative metric of transesterification extent, as each ester bond cleaved via nucleophilic glycol attack produces exactly one terminal hydroxyl group. The exceptional agreement between experimental and theoretical HV values across all systems (Table 4, deviation typically <5%, maximum ± 4.2% for EG 1:0.5) allows us to make an assumption about an approximate, close to near-quantitative conversion.

The proximity of experimental and theoretical hydroxyl values indicates a high degree of reaction completeness and minimal reagent loss during glycolysis. Since the hydroxyl value (expressed in mg KOH/g) is directly proportional to the molar concentration of hydroxyl groups in the sample, matching experimental HV values to those calculated from the targeted oligomer composition implies that almost all hydroxyl functions predicted by the ideal transesterification scheme are actually present and titratable. In the context of PLA glycolysis, each ester bond cleavage by a glycol molecule generates a new terminal hydroxyl group; therefore, any significant deviation between experimental and theoretical HV would reflect either incomplete ester cleavage (residual long chains), competitive side reactions consuming hydroxyl groups, or loss of volatile glycol during processing. The fact that the measured HV values differ from the theoretical ones by less than about 5% across all systems (Table 4) thus provides strong evidence that: (I) the majority of ester linkages accessible under the chosen conditions are cleaved, (II) side reactions that would consume or mask hydroxyl groups are limited, and (III) losses of glycolysis agent are negligible on the scale of the HV determination. This interpretation is consistent with the GPC data, which show a shift in the molecular weight distributions toward the low-molecular-weight region and the absence of significant high-molecular-weight tails for compositions with the highest hydroxyl values.

Within a given glycolysis system, the inverse proportionality between number-average molecular weight and the concentration of chain-end hydroxyl groups means that the experimentally determined HV and the Mn values obtained from GPC are mutually consistent: samples with higher HV systematically exhibit lower Mn, as expected for progressive chain scission without extensive branching or crosslinking.

Propylene glycol demonstrates particularly high reactivity: across the three measured ratios (1:0.25, 1:0.5, 1:1), experimental HV values match theoretical ones within ±3.6%, suggesting minimal side reactions (e.g., ester–ester crosslinking or unproductive chain scission). In contrast, glycerol exhibits larger deviations (up to +6.8% at a 1:2 ratio), which can be attributed to the competing reactivity of its three hydroxyl groups, promoting intramolecular esterification side pathways.

In practical terms, products with HV ≥ 800 mg KOH/g, corresponding approximately to oligomers with degrees of polymerization below about 8, exhibited complete water solubility at 90 °C and residue-free drying. These characteristics are critical for replacing chlorinated solvents in equipment cleaning applications.

3.4. Lactide Synthesis

Based on the concept of secondary production of polylactide from glycolysis products, lactide was synthesized according to the procedure described in Section 2.3. Glycolysis products obtained with glycerol as the glycolysis agent at PLA:GL = 1:0.25 were selected as the starting material for thermocatalytic depolymerization of lactic acid oligomers. Tin octoate at a concentration of 0.5 wt.% relative to the reaction mixture was used as the depolymerization catalyst.

The choice of glycerol at this ratio is justified by the fact that the optimal degree of polymerization of oligomers for lactide production is typically in the range 5–20, which corresponds well to the degree of polymerization of oligomers obtained at PLA:GL = 1:0.25 (Table 2, entry 18) [47].

Glycerol is recognized as safe for humans and is used in cosmetics and pharmaceutical formulations [49]. The high boiling point of glycerol (about 290 °C), which is significantly higher than the temperature of the glycolysis process, helps to minimize the loss of glycolysis agent in the form of vapor. In addition, the presence of hydroxyl-containing glycerol units in lactic acid oligomer chains can increase lactide yield by increasing the number of hydroxyl groups involved in lactide formation reactions [50].

Figure 12 shows the ¹H NMR spectrum of raw lactide obtained from glycolysis products at PLA:GL 1:0.25.

In the spectrum of raw lactide (Figure 12), doublets are observed at 1.42 ppm (signal 2) and 1.50 ppm (signal 2′), which correspond to the methyl protons in L- and D-lactides and in meso-lactide, respectively. The signal at 5.04 ppm (signal 1) is assigned to the methine protons (–CH–) in the lactide ring. The presence of oligomeric impurities is indicated by signals 3, 5, and 4 at 4.25, 1.23, and 1.35 ppm, which are associated with protons at the methine carbon (–CH–), methyl group protons, and methyl protons (–CH₃), respectively, in short lactic acid oligomers.

Signals 6 and 7 in the 4.0–4.2 ppm region are tentatively assigned to the protons at terminal and central carbon atoms in disubstituted glycerol units. This glycerol remains in small amounts in the chains of oligomeric glycolysis products. Signals 8 and 9 at 3.4 and 3.6 ppm correspond to the protons of terminal and central carbon atoms in monosubstituted glycerol. The presence of glycerol ester impurities in the raw lactide can be explained by the fact that, during synthesis, small amounts of modified short-chain lactic acid oligomers are co-sublimed with the lactide vapor and condensed in the receiving flask.

Lactide purification was carried out according to the methodology described in Section 2.3. As can be seen from Figure 13, after five recrystallizations from butyl acetate, the signals associated with meso-lactide and lactic acid oligomers (signals 2′, 3, 4, and 5 in Figure 12) disappear, indicating high lactide purity.

Table 5. Impurity content in raw lactide and purified calculated from ¹H NMR spectra.

№	Name of the Substance	Content, %
		Raw Lactide	Purified Lactide
1	L-lactide	61	99+
2	Meso-lactide	24.4	-
3	Lactic acid oligomers	14.6	less than 1

shows the impurity content in raw lactide and purified calculated from ¹H NMR spectra.

Table 6. Mass yield of lactide after each recrystallization step based on raw lactide after each recrystallization step and comparison with lactide yields from unmodified oligomers.

№	Number of Recrystallizations	Yield from the Initial Mass from Glycolysis Products Lactide, %	Yield from the Initial Mass from Unmodified Oligomers Lactide [2], %
1	1 recrystallization	71.9	71.6
2	2 recrystallization	71.7	55.8
3	3 recrystallization	55.0	51.9
4	4 recrystallization	45.0	41.4
5	5 recrystallization	38.9	-

presents the mass yield of lactide after each recrystallization step based on the initial mass of raw lactide. For comparison, data for lactide purification from unmodified oligomers obtained in our previous work are also included [2].

The recovery of 61% L-lactide in raw material reflects competitive pathways during thermocatalytic depolymerization. The 24.4% meso-lactide formed during the 140–200 °C polycondensation stage indicates partial racemization, consistent with literature reports that lactic acid racemizes in the presence of tin catalysts [2,51,52,53,54]. The cumulative yield of 38.9% compares favorably with indirect hydrolysis-based recovery (35–42% literature range), validating the economic potential of glycolysis- mediated recycling [55].

3.5. Differential Scanning Calorimetry

The purity of lactide was further verified by differential scanning calorimetry. The melting point of the purified lactide was found to be 98.11 °C, which corresponds to the characteristics of pure lactide in the L-form [56]. Figure 14 shows the DSC melting curve of purified lactide sample.

4. Conclusions

In conclusion, the main achievements of this work can be summarized as follows:

• Starting from high-molecular-weight PLA with Mn ≈ 165 kDa, we established a glycolysis protocol that reduces Mn to ≈200–240 g·mol⁻¹ at 220 °C, 60 min and PLA:PG = 1:1, i.e., by almost three orders of magnitude, while maintaining Mw below 300 g·mol⁻¹ and hydroxyl value deviations below 5% relative to theoretical predictions. Under these conditions, propylene-glycol-derived oligomers with Mn < 400 g·mol⁻¹ form homogeneous 10 wt.% aqueous solutions at 90 °C and exhibit low viscosities, providing a practical basis for efficient equipment cleaning with simple water rinsing.
• We demonstrated that glycerol plays a complementary role in PLA glycolysis: at moderate PLA:GL ratios (e.g., 1:0.25), it affords oligomers with Mn ≈ 460 g·mol⁻¹ (degree of polymerization ≈ 5–20) and a high density of hydroxyl groups, which are particularly well suited for subsequent lactide formation. Using these glycerol-based oligomers, we obtained raw lactide containing 61% L-lactide and, after five recrystallizations, >99% pure L-lactide with an overall mass yield of 38.9%, thereby demonstrating a closed-loop route from waste PLA back to high-purity monomer.
• By comparing four structurally diverse polyols, we established practical structure–property guidelines: propylene glycol maximizes depolymerization rate and water solubility of the products, whereas glycerol optimizes oligomer structure and functionality for lactide production, making these two systems particularly attractive for in situ reactor cleaning and valorization of PLA-rich waste streams.
• Compared with hydrolysis, which primarily yields lactic acid, and pyrolysis, which typically requires 260–350 °C, our results show that glycolysis offers a complementary, lower-temperature route that directly furnishes lactide-compatible, water-miscible oligomers and thus simplifies both cleaning and closed-loop recycling schemes.
• Finally, we demonstrated that glycolysis-derived PLA oligomers combine controlled molecular weight with high hydroxyl functionality, making them promising reactive intermediates for advanced materials; by varying the glycolysis agent and PLA:GA ratio, the density and distribution of hydroxyl groups along the chain can be tuned, enabling future design of segmented polyesters and polyurethane-type networks with tailored crosslink density, glass transition temperature, and hydrophilicity.