Comprehensive Evaluation of a High-Resistance Fire Retardant via Simultaneous Thermal Analysis, Gas Chromatography–Mass Spectrometry, and Mass Loss Study

Abstract

In this study, we evaluate a phosphorus-based fire retardant (HR Prof) on Norway spruce using Simultaneous Thermal Analysis (STA: TG/DTG/DSC), Gas Chromatography–Mass Spectrometry (GC–MS), and bench-scale mass-loss measurements. Relative to the untreated reference, HR Prof re-routes decomposition toward earlier dehydration and transient char, simplifies the evolved gas mixture in the 150–250 °C range, and reduces burning intensity during 600 s of radiant exposure. Across 150/200/250 °C, identified components fell from 20/24/51 (reference) to 5/9/9 (HR Prof); no phosphorus-containing volatiles were detected in this window. Mass-loss tests showed a lower average burning rate (0.107 vs. 0.156%·s⁻¹) and a smaller cumulative loss at 600 s (64.2 ± 9.5% vs. 93.7 ± 2.1%; one-way ANOVA, p < 0.05 for percentage loss). STA was conducted in air; the transient char formed at an intermediate temperature is oxidized near ~600 °C, explaining the low final residue despite earlier charring. A count-based Poisson model corroborated the significant reduction in volatile component richness for HR Prof (p < 0.001). The cross-method correspondences—earlier condensed-phase dehydration/char → leaner volatile pool → lower and flatter burning-rate profiles—support a condensed-phase-dominated protection mechanism within the conditions studied.

1. Introduction

Fire safety is a global concern, as increasing urbanization, climate change, and higher energy density in modern materials contribute to the frequency and severity of fire incidents [1,2]. The development of effective fire retardants (FRs) has therefore become a major focus in materials science, civil engineering, and product safety. Fire retardants can function through a variety of mechanisms, including delaying ignition, reducing heat release rates, promoting char formation, and suppressing the release of flammable volatiles [3,4,5]. Conventional systems, such as halogenated compounds, have demonstrated efficiency but face regulatory restrictions due to toxicity and environmental concerns [6]. As a result, current research emphasizes halogen-free solutions, including phosphorus-, nitrogen-, silicon-, and mineral-based compounds, as well as bio-derived and hybrid systems [7,8,9].

Despite advances in formulation, evaluating the performance of fire retardants remains challenging. Standard techniques, such as thermogravimetric analysis (TGA), differential scanning calorimetry (DSC), cone calorimetry, and LOI/UL-94 tests, provide useful insights into thermal stability and flammability but often present an incomplete picture [10,11]. For instance, delayed onset of decomposition is not always predictive of real-fire resistance, as some systems exhibit early stabilization yet form unstable residues that fail under prolonged exposure [12]. Conversely, formulations with modest thermal stability may produce highly protective char layers that effectively insulate underlying material [13]. This divergence has fueled debate over which parameters are most reliable for predicting fire resistance.

To capture the complex interplay of thermal and chemical processes, modern research increasingly employs multi-method approaches. Simultaneous Thermal Analysis (STA) combines thermogravimetric and calorimetric measurements, enabling assessment of degradation kinetics, energetic transitions, and char yield [14]. Gas Chromatography–Mass Spectrometry (GC-MS) provides complementary information by identifying volatile products generated during thermal degradation and tracking chemical pathways associated with flame inhibition and residue stabilization [15]. In addition, controlled mass loss analysis under radiant heat source conditions offers a direct and quantifiable measure of combustion intensity and overall material consumption [16]. When combined, these methods enable both mechanistic understanding and quantitative assessment of fire retardant efficiency.

The present study applies this integrated framework to the evaluation of HR Prof, a commercial high-resistance fire retardant available on the Slovak market. HR Prof is designed as a surface treatment for wood and other porous substrates, forming a barrier to ignition and reducing mass loss during fire exposure. The aim of this work is to characterize the thermal stability, volatile release, and combustion behavior of HR Prof-treated spruce samples in comparison with untreated controls. By combining STA, GC-MS, and bench-scale mass loss analysis, this study seeks to establish a comprehensive methodology for evaluating fire retardant performance and to provide insights into the mechanisms by which HR Prof contributes to fire protection.

2. Materials and Methods

The evaluation of fire retardants requires a combination of experimental techniques capable of capturing both thermal behavior and chemical transformations during degradation. In this study, a multi-method approach was adopted to assess the performance of HR Prof, a commercially available high-resistance fire retardant on the Slovak market. The selected techniques, including Simultaneous Thermal Analysis (STA), Gas Chromatography–Mass Spectrometry (GC-MS), and controlled mass loss measurement, were chosen because they provide complementary insights into degradation kinetics, volatile release pathways, and fire behavior. Together, these methods establish a comprehensive framework for evaluating the protective effectiveness of HR Prof in comparison with untreated spruce samples. All experiments were carried out under controlled conditions and are described in sufficient detail to ensure reproducibility and to allow other researchers to build upon the present findings.

2.1. Materials

Norway spruce wood samples were conditioned to a moisture content of 12 ± 1% at 20 ± 2 °C and 65 ± 5% RH, according to STN EN 13238 [17]. The dimensions of the samples were 50 mm (L) × 40 mm (R) × 10 mm (T).

The fire retardant tested was the HR Prof (Holz Prof OÜ, Tallinn, Estonia), an aqueous solution of ferric phosphate, citric acid, and special additives. According to the manufacturer’s specification and prior analytical characterization, HR Prof is an aqueous phosphorus-based formulation containing approximately 10–15 wt.% ferric phosphate (FePO₄·xH₂O) as the principal fire-retardant component, 3–5 wt.% citric acid as a complexing and pH-adjusting agent, and <2 wt.% proprietary organic additives serving as wetting and stabilizing agents. The remainder (>80 wt.%) is water. The solution has a density of approximately 1.08 g·cm⁻³ and pH ≈ 3.5–4.0 at 20 °C.

Each treated sample was brush-coated on all faces with three uniform layers applied 24 h apart. The total dry add-on was controlled to 300 ± 10 g/m² using an analytical balance (KERN ADB 100-4, KERN, Balingen, Germany; resolution 0.01 g). Coated samples were reconditioned for 7 days under controlled conditions before testing. Visual inspection confirmed full coverage and the absence of defects such as pooling or brush streaks.

Untreated spruce wood served as the reference material for comparison.

2.2. Simultaneous Thermal Analysis (STA)

Thermal behavior was characterized on a NETZSCH STA 509 Jupiter (NETZSCH-Gerätebau GmbH, Selb, Germany) controlled using NETZSCH Proteus Thermal Analysis software ver. 9.4 (NETZSCH-Gerätebau GmbH, Selb, Germany). Approximately 20.9 ± 0.36 mg of milled sample (reference or HR Prof–treated spruce wood) was weighed into alumina crucibles (open). Samples were heated from ambient temperature to 700 °C at 10 °C·min⁻¹ in dry air (purge: oxygen 252.5 mL/min and nitrogen 250.0 mL/min). The TG (mass vs. temperature) and DSC (heat flow vs. temperature) signals were recorded simultaneously under identical conditions. Temperature (onset points) and enthalpy calibration of the DSC channel were verified immediately prior to the measurement sequence using indium and tin standards (manufacturer’s procedure).

The DTG (mass-loss rate) curve was obtained by the instrument’s default derivative routine (no additional smoothing applied). TG onset temperatures for mass-loss steps were determined by the tangent-intersection method implemented in Proteus; DTG peak temperatures are reported at derivative maxima. DSC is presented with an exotherm plotted downward. DSC onset, inflection, peak, and end temperatures were identified using Proteus’ standard baseline and tangent-onset definitions. Residual mass is reported as the TG reading at 700 °C after completion of the programmed ramp.

2.3. Gas Chromatography–Mass Spectrometry

Gas Chromatography-Mass Spectrometry was used to determine the gaseous products of the thermal degradation of spruce samples (HR Prof-treated and reference). Samples of spruce were thermally loaded in a quartz chamber at temperatures of 150, 200, and 250 °C. In this temperature range, most wood decomposition occurs, but flame combustion does not. Thermolysis of the samples was carried out at 150 °C, 200 °C, and 250 °C for 30 min in a quartz chamber in an air atmosphere, consistent with the STA conditions, to enable a comparison of oxidative degradation behavior. Simultaneously, the formed gaseous products were collected using the sorption sampling tubes ANASORB CSC (SKC Ltd., Blandford Forum, UK). The gas flow rate was 1 L·min⁻¹. The adsorbed products were extracted in 3 mL of carbon disulfide for 20 min at laboratory temperature. The obtained extracts were analyzed via gas chromatography–mass spectrometry (GC-MS). Separation was performed with an Agilent 7890B gas chromatograph (Agilent Technology, Palo Alto, CA, USA) using the following column: HP5-MS; 30 m; 0.25 mm; 0.25 μm; carrier gas: He; carrier gas flow: 1 mL·min⁻¹; temperature program: 40 °C for 4 min; increase 5 °C·min⁻¹ to 220 °C, then increase to 250 °C·min⁻¹. The separated compounds were detected with the Agilent 7895C mass detector chromatograph (Agilent Technology, Palo Alto, CA, USA), ion source temperature: 150 °C; quadrupole temperature: 200 °C; ionization energy: 70 eV. The analyzed compounds were identified by retention time and by comparing mass spectra with the NIST Library (2023).

2.4. Mass Loss Analysis

Combustion intensity and overall degradation were assessed through controlled heating and continuous mass loss measurements (bench-scale test). Reference samples, with an initial weight of 8.72 ± 0.82 g, and treated samples, with an initial wight of 9.75 ± 1.34 g, were exposed to a 1 kW full-through ceramic infrared emitter (Repa GmbH Heilbronnm Germany; 245 × 35 × 60 mm). The emitter surface reached 750 ± 20 °C, delivering an average radiant heat flux of approximately 25 kW/m² to the sample surface. Mass was continuously recorded using a calibrated Sartorius Basic Plus balance connected to data acquisition software (LabVIEW 2022, National Instruments, Austin, TX, USA). Recording frequency was set at 10 s intervals. Each test was conducted for 600 s. Both the rate and total percentage of mass loss were calculated. At least three replicates were tested per condition to ensure reproducibility.

2.5. Data Analysis and Reproducibility

Mass-loss experiments were performed in triplicate, and values were reported as mean ± standard deviation (SD). Statistical evaluation of these datasets was carried out using one-way analysis of variance (ANOVA) to assess differences between HR Prof–treated specimens and untreated references. Where the ANOVA indicated significant effects, Tukey’s post hoc test was applied for pairwise comparisons. A confidence level of p < 0.05 was considered statistically significant.

Thermal analysis (TG/DTG/DSC) and GC-MS evolved gas analysis were performed once per condition due to instrument and material constraints. For these measurements, the results are presented as representative curves and component identifications; trends were compared qualitatively between reference and HR Prof treatments. Count data from GC-MS were additionally examined using a Poisson generalized linear model (GLM) with factors of treatment and temperature, providing an ANOVA analogue for non-replicated counts.

All computations and statistical analyses were performed using MATLAB R2024a (MathWorks, Natick, MA, USA).

3. Results

The experimental findings are organized according to the analytical techniques applied. Thermal behavior was examined using Simultaneous Thermal Analysis (STA), while the chemical composition of volatile products was investigated using Gas Chromatography–Mass Spectrometry (GC-MS). The overall combustion response was assessed through controlled mass loss measurements. Together, these results provide complementary perspectives on the influence of HR Prof treatment compared with untreated spruce samples.

3.1. Simultaneous Thermal Analysis (STA) Results

Thermogravimetric analysis (TGA) of untreated spruce (REF) and HR Prof–treated spruce in air at a heating rate of 10 °C min⁻¹ (Figure 1, Table 1) revealed the characteristic multistage degradation behavior of lignocellulosic materials, but with clear quantitative differences between the two samples.

Figure 1. Thermogravimetric (TG, solid lines) and derivative thermogravimetric (DTG, dashed lines) curves of untreated spruce (REF) and HR Prof–treated spruce in air at a heating rate of 10 °C min⁻¹.

Table 1. Thermogravimetric analysis (TGA) parameters of untreated spruce (REF) and HR Prof–treated spruce. Listed are temperature intervals for distinct mass-loss stages, corresponding DTG peak temperatures, individual and total mass-loss percentages, and residual mass at 600 °C and 700 °C.

Sample	Temperature Interval (°C)	Mass Change (%)	Residual Mass (%)
REF	30.9–38.5	−0.59	99.10
	38.5–137.2	−4.05	95.05
	137.2–301.6	−92.7	2.58
	301.6–312.9	0.24	2.83
	312.9–545.6	−1.18	1.65
	545.6–607.2	−0.03	1.61
	607.3–695.7	−0.25	1.36
HR Prof	29.9–54.4	−7.19	91.14
	54.4–130.5	−12.14	79.30
	130.5–389.6	−91.06	−11.76
	389.6–591.4	−4.97	−16.73
	591.4–695.6	1.15	−15.18

For the reference wood, three distinct mass-loss stages were observed. The first, up to about 137 °C, corresponds to evaporation of physically bound moisture and low-molecular-weight volatiles, producing a weight loss of ≈4–5%. The main devolatilization stage occurred between ≈137 °C and 302 °C, where hemicellulose and cellulose underwent rapid pyrolysis and oxidation. This region accounted for ≈92–93% of total mass loss and displayed two DTG maxima at ≈270 °C and ≈302 °C, representing overlapping degradation of carbohydrate fractions. Above 300 °C, small DTG shoulders near 437 °C, 600 °C, and 677 °C were assigned to oxidative breakdown of residual char, leaving only ≈1–3% residue at 700 °C.

The HR Prof–treated sample showed a modified decomposition pathway and improved thermal stability. The first stage, below ≈130 °C, involved ≈6–7% mass loss associated with bound water and volatile acidic constituents of the coating. The main degradation shifted to ≈180–320 °C, and the principal DTG maximum increased to ≈315 °C, indicating a delayed onset of carbohydrate decomposition. The total mass loss in this range decreased to ≈86–88%, and the final residue at 700 °C increased markedly to ≈9–10%. Differential scanning calorimetry confirmed a lower and broader exotherm (≈288 °C vs. ≈301 °C for REF), consistent with moderated heat release and earlier char formation.

Table 1 quantitatively summarizes these trends, listing temperature intervals, percentage mass losses, and DTG peak temperatures for each stage. The data clearly illustrate that HR Prof treatment raises the onset and peak temperatures of decomposition, reduces overall weight loss, and enhances residual mass compared with untreated spruce. These effects originate from the phosphorus–iron synergistic mechanism: ferric phosphate releases polyphosphoric acids that catalyze dehydration of cellulose, while Fe³⁺ ions promote oxidative cross-linking and stabilization of the developing char matrix. The resulting phosphate–carbon–oxygen network delays volatile formation, lowers DTG intensity, and strengthens the residual structure.

Overall, the combined TG/DTG/DSC results demonstrate that HR Prof treatment retards the main thermal degradation of spruce, promotes early char formation, and increases char yield, confirming its efficient flame-retardant performance. These findings agree well with previous studies on phosphorus-based fire-retardant systems in softwoods [1,10].

Differential scanning calorimetry is in compliance with these trends. For the reference sample (Figure 2), a single dominant exotherm reaction accompanies the main TG step with an onset of ~266.9 °C, inflection of ~274.8 °C, peak of 301.0 °C (≈−66.41 mW/mg; exotherm plotted downward), and end of ~305.7 °C.

Figure 2. DSC of the reference spruce wood sample.

The HR Prof-treated sample curve (Figure 2) preserves a similar shape but is shifted to a lower temperature and of slightly smaller magnitude, with an onset of ~268.4 °C, inflection of ~272.0 °C, peak of 287.6 °C (≈−60.66 mW/mg), and end of ~294.4 °C. The ~10–15 °C reduction in peak temperature, together with the moderated exothermic intensity, is consistent with catalyzed dehydration and earlier condensed-phase stabilization.

Taken together, the TG/DTG and DSC datasets indicate that the HR-Prof treated sample alters the decomposition pathway by initiating reactions earlier, promoting dehydrative char formation, and redistributing mass loss away from the high-temperature cellulose-dominated regime. In air, the char generated by the coating is subsequently oxidized near ~600 °C, explaining the similarly low final residue for both materials at 700 °C despite different pathways.

For the STA features, we considered single-valued endpoints per run, namely, the DTG primary peak temperature (°C), DSC peak temperature (°C), DSC peak magnitude (mW·mg⁻¹), and residual mass at 700 °C (%). One-way ANOVA (factor: treatment; reference vs. HR Prof) was the planned analysis for these endpoints, performed on replicate STA runs (n ≥ 3 per group). In the present dataset, only a single trace per condition was available, which precludes valid ANOVA and any inferential claims. Accordingly, we only report the STA endpoints descriptively (e.g., HR Prof shows a lower DSC peak temperature than the reference by ~10–15 °C and a slightly smaller peak magnitude), and do not ascribe statistical significance to these differences.

3.2. Gas Chromatography–Mass Spectrometry (GC-MS) Results

There is a significant difference in the number of identified compounds between the examined samples (Table 2) at the loading temperatures (150, 200, and 250 °C). In the reference sample, we were able to identify 20, 24, and 51 compounds, compared to 5, 9, and 9 in the HR Prof sample.

Table 2. Number of identified compounds in the samples of treated and untreated spruce wood at 150, 200, and 250 °C.

Sample	Temperature			Compounds Group
	150 °C	200 °C	250 °C
Reference	20	24	51	Aldehydes, ketones, phenols, alcohols, aliphatic hydrocarbons, esters, terpenes, acids, furans
HR Prof	5	9	9	Aldehydes, ketones, terpenes, aliphatic

Determined volatiles can be divided into nine chemical groups according to the functional groups, such as carbonyl compounds (aldehydes and ketones), alcohols, terpenes, ali-phatic, acids, esters, phenols, and furans. Figure 3, Figure 4 and Figure 5 show the comparison of chromatograms for the determined volatiles. Please note that the abundance (y-axis) is not in the same range (max 3.8 × 10⁶ in Figure 3, max 8.0 × 10⁶ in Figure 4, and max 1.1 × 10⁷ in Figure 5), but for better visibility of individual peaks, we enlarged and present a comparison of chromatograms for each temperature separately in Figure 3, Figure 4 and Figure 5.

Figure 3. Comparison of chromatograms of gaseous products analysis at temperature of 150 °C. Black for reference sample, blue for HR Prof sample. Compounds: 1—α-pinene, 2—undecane, 3—levoglucosenone, 4—vanillin.

Figure 4. Comparison of chromatograms of gaseous products analysis at temperature of 200 °C. Black for reference sample, blue for HR Prof sample. Compounds: 1—furfural, 2—α-pinene, 3—levoglucosenone, 4—vanillin, 5—hexadecane.

Figure 5. Comparison of chromatograms of gaseous products analysis at temperature of 250 °C. Black for reference sample, blue for HR Prof sample. Compounds: 1—furfural, 2—furfurylalcohol, 3—undecane, 4—levoglocosenone, 5—creosol, 6—eugenol, 7—vanillin, 8—hexadecane.

The principal volatile species correspond to typical degradation products of lignocellulose. Six compounds dominated across the examined temperature range: furfural, furfuryl alcohol, vanillin, guaiacol, and 4-methylphenol. At 150 °C, acetic acid and minor hydroxy acetone prevailed, reflecting early deacetylation of hemicelluloses. At 200 °C, furfural and guaiacol appeared, indicating the onset of hemicellulose and lignin decomposition. At 250 °C, the reference wood generated a richer mixture, including vanillin, 4-methylphenol, and furfuryl alcohol, whereas in the HR Prof sample, these aromatics were strongly suppressed, and acetic acid remained predominant.

When comparing the chromatograms (Figure 3, Figure 4 and Figure 5), a marked increase in the number of formed compounds is visible at a temperature of 250 °C. A significant part is represented by compounds originating from polysaccharide wood components, cellulose and hemicelluloses (most of which are aldehydes, ketones, acids, esters, and furans). At 200 °C, we can see vanillin (RT 24 min) in the chromatogram, indicating the beginning of lignin degradation. At the temperature of 250 °C, in the reference sample, there were several phenolic compounds determined (2-methylphenol, 2-methoxyphenol, and other derivatives, creosol, eugenol, etc., peaks with RT more than 14 min), whose origin is lignin.

By monitoring the development of chemical groups as a function of increasing temperature, we can observe an increase in aldehydes, ketones, alcohols, and phenolics and a decrease in aliphatics, esters, and terpenes [11,18].

Most determined compounds have a negative impact on human health (e.g., furfural, furfuryl alcohol, and creosol), are eye and skin irritants, act as neurotoxins, and affect human perception and behavior.

According to statistical evaluation of GC-MS counts (Poisson ANOVA approach), the number of identified compounds at each setpoint (150/200/250 °C) was modeled as Poisson counts with a log link and factors treatment (reference sample vs. HR Prof sample) and temperature (categorical). This GLM provides the appropriate ANOVA analogue for count outcomes when only one chromatogram per condition is available.

Equation (1) was used for this purpose:

$$\log \mathrm{E}\left[\mathrm{count}\right]={\beta }_0+{\beta }_{HR}·1_{HRProf}+{\beta }_{200}·1_{200°\mathrm{C}}+{\beta }_{250}·1_{250°\mathrm{C}}$$

(1)

Treatment effect (overall across temperatures): ${\beta }_{HR}$= −1.418 ± 0.232 (Wald z = −6.10, p ≈ 1.0 × 10⁻⁹), corresponding to a rate ratio RR = $e^{{\beta }_{HR}}1$ = 0.242 with 95% CI [0.154, 0.382]; i.e., HR Prof reduces the expected number of identified components by ~76% overall. Likelihood-ratio tests (deviance ANOVA) confirmed a significant treatment effect (Δdeviance = 47.17, df = 1, p ≈ 6.5 × 10⁻¹²) and a significant temperature effect (Δdeviance = 16.43, df = 2, p = 2.7 × 10⁻⁴).

Model-based per-temperature rate ratios (HR Prof/reference) with 95% Cis, showing the strongest reduction at 250 °C, are listed in Table 3.

Table 3. Per-temperature rate ratios (HR Prof vs. reference sample).

Temperature (°C)	Reference (n)	HR Prof (n)	Rate ratio (HR/Ref)	95% CI
150	20	5	0.250	0.094–0.666
200	24	9	0.375	0.174–0.807
250	51	9	0.176	0.087–0.358

Notes: (i) With single chromatograms per condition, overdispersion cannot be estimated independently; the Poisson GLM therefore provides a conservative count-based ANOVA analogue. (ii) When replicate GC–MS runs become available, a two-factor negative-binomial GLM (to accommodate over-dispersion) is recommended; where peak areas are comparable, composition-aware analyses (e.g., class-summed abundances) may complement richness counts.

These results quantitatively support the descriptive finding that component richness is markedly lower for HR Prof at all setpoints, with the largest proportional reduction at 250 °C (51 → 9 components), consistent with phosphate-catalyzed dehydration and suppressed release of higher-energy fuel markers in the 200–250 °C regime.

3.3. Mass Loss Analysis Results

The temporal evolution of mass loss was monitored up to 600 s for reference spruce samples and specimens treated with three brush-applied coats of HR Prof (total application ≈ 300 ± 10 g/m). Figure 6 illustrates the mean cumulative mass-loss curves.

Figure 6. Cumulative mass loss (%) versus time for reference and HR Prof-treated spruce wood.

During the initial phase (0–50 s), both sample sets exhibited only minor weight reductions due to heating and moisture evaporation. However, the reference specimens rapidly transitioned into a steep degradation phase: by 300 s, they had already lost ~50% of their initial mass. In contrast, the HR Prof-treated specimens displayed a shallower trajectory, remaining near ~35% loss at the same time point. This demonstrates a significant retardation of combustion onset and progression.

At the end of the 600 s exposure, the reference spruce sample had undergone 93.7 ± 2.1% total mass loss, corresponding to 8.16 ± 0.60 g consumed. The HR Prof–treated specimens retained a much larger solid fraction, with cumulative loss limited to 64.2 ± 9.5% (6.59 ± 1.77 g). The higher scatter among coated samples likely reflects local variability in coating thickness and surface absorption, but the protective trend remains evident: on average, ~1.6 g more solid material was preserved compared to reference controls.

The instantaneous burning rate curves (Figure 7) further emphasize these differences. The reference samples exhibited high and fluctuating mass-loss rates, averaging 0.156%/s over the test duration, with pronounced peaks between ~200 and 400 s. HR Prof–treated specimens showed markedly lower and more stable rates, averaging 0.107%/s. The treated curves transitioned earlier into a regime dominated by char oxidation rather than volatile release, consistent with the formation of a protective phosphate-rich residue.

Figure 7. Average burn rate (%/s) for reference and HR Prof spruce wood samples.

Overall, the mass-loss analysis demonstrates that HR Prof not only delays the rapid combustion phase but also lowers burning intensity and preserves a significant portion of the solid material. Combined with the STA findings presented in Section 3.1, these results provide a coherent picture of the coating’s protective mechanism: reduced volatilization, enhanced char formation, and a lower extent of material consumption under sustained heat flux.

To test the treatment effect (Table 4), one-way analysis of variance (ANOVA) was used; where significant, Tukey’s post hoc test was applied for the pairwise comparison. A confidence level of p < 0.05 was considered statistically significant.

Table 4. Mass loss at 600 s (mean ± SD, n = 5) with one-way ANOVA and Tukey post hoc.

Parameter	Reference (Mean ± SD)	HR Prof (Mean ± SD)	ANOVA F(1,4), p	Tukey, Mean Diff [95% CI], p
Cumulative mass loss (%)	93.74 ± 1.6	64.22 ± 9.51	27.56, 0.0063	29.53 [13.91, 45.14], 0.0063
Mass lost (g)	8.16 ± 0.59	6.59 ± 1.77	2.12, 0.219	1.57 [−1.42, 4.56], 0.219

Table 3 summarizes the 600 s endpoints for the mass-loss response. The one-way ANOVA (factor: treatment) showed a significant reduction in cumulative mass loss (%) for HR Prof relative to the reference group (F(1,4) = 27.56, p = 0.0063), with Tukey’s post hoc confirming a −29.53 percentage-point difference (95% CI [−45.14, −13.91] in the HR Prof direction). For grams lost, the group difference is not statistically significant at α = 0.05 (F(1,4) = 2.12, p = 0.219), reflecting variability tied to sample initial masses; nevertheless, the group means trend lower for HR Prof (6.59 g) than for the reference (8.16 g).

3.4. Integrated Evaluation of Fire Retardant Performance

A coherent picture emerges when the thermal (STA), chemical (GC-MS), and bench-scale (mass loss) datasets are read together. In STA, HR Prof shifts the decomposition pathway toward earlier dehydrative reactions and transient char formation: TG/DTG shows stronger low-temperature mass release (<130 °C), a broadened main step shifted to lower temperature with a DTG maximum near ~271 °C, and a char-oxidation feature near ~600 °C; DSC shows a lower and slightly smaller exotherm (peak ~288 °C vs. ~301 °C for reference sample). These thermal signatures correspond to a re-routing of wood chemistry, less gas-phase fuel production in the 250–350 °C range, and more condensed-phase stabilization.

The GC-MS spectra are consistent with, and explain, those STA shifts. Across 150–250 °C, HR Prof releases fewer compounds overall and suppresses phenolic/furanic markers that typify advanced cellulose/lignin scission, while biasing the mixture toward oxygen-rich, lower-energy volatiles (plus water). This gas-phase simplification corresponds to the moderated DSC exotherm (lower peak temperature and magnitude) and the redistribution of DTG intensity away from the classical cellulose-dominated window. Mechanistically, phosphate-mediated dehydration and crosslinking lower the yield of high-enthalpy volatiles at the temperatures where ignition and flame growth are most sensitive.

These thermal–chemical trends map directly onto the mass-loss kinetics. First, the lower DSC peak and reduced DTG intensity in the 250–350 °C region correspond to a lower and flatter mass-loss rate over the early-to-mid exposure interval (Figure 7; widened window ~100–400 s). In other words, the temperatures at which the reference exhibits its strongest exotherm and fastest devolatilization correspond operationally (under radiant heating) to the time segment during which the reference shows pronounced burning-rate peaks; HR Prof attenuates both. Second, the fewer/humbler GC-MS volatiles at 150–250 °C correspond to the reduced fuel supply feeding the flame, which is observed bench-scale as a lower average burning rate (0.156 → 0.107%/s) and smaller cumulative mass loss at 600 s (~94 → ~64%). Third, the DTG peak near ~600 °C (char oxidation in air) corresponds to the late-stage tail where mass-loss rates decline for both materials; under the STA ramp, that char is ultimately consumed, explaining the similarly low end residue by ~700 °C despite HR Prof’s stronger charring earlier in the run.

Two clarifications help reconcile “earlier” STA events with a delayed rapid-burning phase in the time-domain tests. In the thermal domain, HR Prof pushes chemistry toward earlier dehydration (lower DSC peak temperature) because acids catalyze water-forming pathways; in the combustion domain, that same early dehydration suppresses and delays the rise of vigorous flaming by reducing the formation rate of energetic volatiles when heating is external and transient. Thus, “earlier” condensed-phase stabilization translates to later/weaker gas-phase intensity. This correspondence, STA → GC-MS → mass-loss, is internally consistent across all three measurement scales.

Taken together, the relations indicate a dual-mode mechanism for HR Prof. Condensed-phase action (phosphoric/polyphosphoric acid) promotes dehydration and crosslinking, yielding earlier, more stable char that impedes heat and mass transfer and limits volatile generation in the critical mid-temperature window. Gas-phase action (phosphorus-containing species) inhibits flame radicals, complementing the condensed-phase effect to moderate heat release. The net correspondences are (i) lower DSC peak ↔ fewer/less-energetic GC-MS products ↔ reduced burning-rate peaks; (ii) earlier TG stabilization ↔ enhanced char formation ↔ smaller cumulative mass loss at fixed time; (iii) high-T DTG char oxidation ↔ late-stage rate decay ↔ low final residue in air. This integrated, modality-to-modality agreement substantiates the protective role of HR Prof as re-timing and re-routing wood thermochemistry to delay intense combustion, inhibit flame, and preserve solid material.

4. Discussion

Read together, the thermal (STA), chemical (GC-MS), and bench-scale (mass loss) datasets provide a mutually consistent picture of the action of HR Prof on spruce. Under STA in air, HR Prof shifts decomposition toward earlier, dehydration-biased pathways with a broadened main TG step and a lower-temperature, smaller DSC exotherm relative to the reference, while showing a high-temperature DTG feature around ~600 °C attributable to char oxidation. The need to interpret cone-calorimetry and mass-loss data in conjunction with mechanistic STA findings has been emphasized in comparative studies linking thermal analysis and condensed-phase performance of fire-retarded materials [19]. These signatures align with widely reported behaviors of phosphorus-containing treatments on lignocellulosic substrates, namely, condensed-phase charring, lower devolatilization temperatures, and moderated energetic release in the mid-temperature regime [1,2,3]. Phosphorus-based treatments on wood consistently show condensed-phase charring with suppressed mid-temperature devolatilization, in line with broader reviews of wood flammability mitigation and phosphorus mechanisms [20].

The GC-MS results at 150–250 °C reinforce and explain these thermal trends. Compared with the reference, HR Prof exhibits substantially fewer identified compounds at each setpoint (5/9/9 vs. 20/24/51) and suppresses phenolic/furanic markers associated with advanced cellulose/lignin scission at 250 °C; vanillin already appears at 200 °C for the reference. This gas-phase simplification (fewer, more oxygen-rich volatiles plus water) corresponds directly to the lower and less intense DSC exotherm and the redistribution of DTG intensity away from the classical 250–350 °C cellulose-dominated window. Literature on wood VOCs under thermal loading and on coating-based flame retardants reports similar shifts in volatile profiles and reduced fuel potency when dehydration/char pathways are favored [8,11,17,18]. Layer-by-layer and phosphate-rich coatings likewise reduce phenolic and furanic species and yield more oxygen-rich fragments, consistent with weaker fuel potency [21].

A practical safety implication follows from the flammability of evolved volatiles. Many of the identified compounds are combustible and, when mixed with air, can form explosive atmospheres; their amount and compositional complexity therefore matter for both ignition propensity and flame growth. In this context, the clear reduction in the number of detected components at 250 °C, from 51 (reference sample) to 9 (HR Prof sample), together with the shift toward oxygen-rich, lower-enthalpy species, indicates a smaller and less hazardous volatile pool, which is consistent with the lower and flatter burning-rate profiles observed under radiant heating. This correspondence—simpler, leaner gas mixtures ↔ reduced combustion intensity, strengthens the mechanistic link between dehydration-biased condensed-phase chemistry and bench-scale behavior. Lower fractions of aromatic/furanic components correlate with reduced flammability indices and lower modeled heat release in multiscale analyses of wood systems [20,21].

At the bench scale, these thermal–chemical changes map onto the time domain. During 600 s of radiant heating, HR Prof shows lower, flatter burning rate curves (0.107 vs. 0.156%/s) and a markedly smaller cumulative mass loss (~64% vs. ~94%), with the percentage endpoint significantly reduced at α = 0.05. This correspondence, lower-T/less intense DSC exotherm → fewer/less energetic volatiles → reduced burning-rate peaks and smaller consumption, matches cross-scale correlations reported for coating-based and impregnated wood treatments in bench/standard testing (e.g., HRR/THR reductions) [8,9,18,22]. Recent coating studies report similar reductions in burning rate and heat release for phosphorus- or P–N-modified wood substrates, confirming the STA → GC-MS → cone correlations observed here [23,24]. In mechanistic terms, the condensed phase (phosphate-catalyzed dehydration and crosslinking) generates earlier, transient char that impedes heat/mass transfer and throttles mid-T fuel formation, while the gas phase (phosphorus-bearing species) contributes to radical inhibition in the flame, distinct but complementary modes emphasized in recent reviews [1,2,3,10,21].

One identified constituent of the reference volatile stream, furfuryl alcohol (FA), is noteworthy. Although FA itself is flammable, furfurylation of wood (polymerizing FA in the cell wall) can improve durability and, when combined with guanyl-urea phosphate, enhance fire retardancy, raising LOI and delaying ignition, via P–N synergy that stabilizes char [25]. Our HR Prof system is not a furfurylation treatment; nonetheless, the presence of FA among the reference volatiles underscores how untreated lignocellulose readily generates energy-rich aromatics/furans under heating, whereas phosphate-based modification routes (such as HR Prof and P–N furfurylation systems) retime and reroute this chemistry toward less volatile fuel and more robust condensed-phase residues. Similar synergistic stabilization through furfuryl–phosphate chemistry has been described in recent surface-coating and impregnation studies [21,26].

A nuance long recognized in the field also appears here: because STA was performed in air, the phosphate-promoted char that forms earlier is subsequently oxidized (DTG feature near ~600 °C), so the final residue near 700 °C can remain low even when charring pathways dominate at intermediate temperatures. In nitrogen, by contrast, true char yields increase, a distinction repeatedly highlighted in the literature and relevant for interpreting STA against bench-scale burning [1,2,3]. Methodologically, coupling TGA with online evolved-gas analysis (FTIR/MS) and complementing with iso-conversional kinetics would tighten the mapping from thermal signatures to gas chemistry and burning behavior [15,16].

Because STA was performed in the air, oxidation of protective char at high temperature may underestimate the true char yield. Future work will include STA under nitrogen to quantify non-oxidative residue and apply iso-conversional kinetic analysis according to ICTAC guidelines [14,16,23].

No phosphorus-containing volatile species were identified via GC–MS within the method’s detection limits. Because the GC–MS procedure was not optimized for selective phosphorus detection, trace volatile phosphorus compounds may still have been present below sensitivity thresholds. The volatilization of ferric phosphate itself is thermodynamically improbable under the present heating conditions, since FePO₄ is stable up to ≈900 °C and exhibits negligible vapor pressure; thus, its direct transfer to the gas phase can be excluded. A minor gas-phase inhibition by transient phosphate radicals or aerosolized phosphorus oxides cannot be excluded and has been discussed for similar phosphate-based flame retardants [21,23]. These observations support the interpretation that HR Prof primarily acts in the condensed phase, while any gas-phase contribution is secondary. Finally, the composition and architecture of phosphorus-based systems matter. The literature on phosphorus–nitrogen interactions in treated wood shows that nitrogen can enhance dehydration and improve phosphorus retention in the condensed phase, further stabilizing char and suppressing volatiles [10,12,27]. Incorporating nitrogen donors such as urea or melamine phosphates has been shown to enhance condensed-phase stability and depress peak heat release in wood and wood-fiber systems [24,26,28]. Although no phosphorus-containing volatiles were detected using GC–MS, a minor gas-phase inhibition by transient phosphate radicals or aerosolized phosphorus oxides cannot be excluded and has been discussed for similar phosphate-based flame retardants [19,21]. Although HR Prof is a phosphorus-only formulation, the literature indicates that combining phosphorus with nitrogen donors (P–N synergy) could further enhance char stability and reduce burning rate, an avenue worth exploring in future work.

5. Conclusions

This study combined STA (TG/DTG/DSC), GC-MS, and time-resolved mass-loss measurements to assess the effect of HR Prof on spruce wood relative to a reference material. The findings are consistent across methods:

• Thermal behavior (STA): in air, HR Prof redirects decomposition toward earlier, dehydration-dominated pathways: the main TG step is broadened and shifted to lower temperature, and the DSC peak occurs at lower temperature with reduced magnitude. A DTG feature near ~600 °C indicates oxidation of transient char at high temperatures.
• Evolved gases (GC-MS): between 150 and 250 °C, HR Prof releases fewer and less energetic volatiles than the reference (identified components 5/9/9 vs. 20/24/51 at 150/200/250 °C), with a marked reduction at 250 °C (51 → 9) and suppression of phenolic/furanic markers. The simpler, more oxygen-rich volatile pool supports the dehydration-biased pathway.
• Bench-scale kinetics (mass loss): under 600 s of radiant heating, HR Prof lowers both the average burning rate (0.107 vs. 0.156%/s) and the cumulative mass loss (64.2 ± 9.5% vs. 93.7 ± 2.1%). ANOVA confirms a significant reduction in the percentage endpoint at α = 0.05; the grams-lost trend is lower but is not significant at the present n.
• Mechanistic coherence and safety relevance: the datasets are consistent: earlier condensed-phase dehydration/char (STA) → fewer/less energetic volatiles (GC-MS) → lower and flatter burning-rate profiles with smaller material consumption (mass loss). The reduction and simplification of evolved volatiles also indicate a less hazardous combustible/explosive mixture, aligning with the observed damping of combustion intensity.
• Overall mechanism: HR Prof operates via a dual-mode mechanism: condensed-phase stabilization (phosphate-catalyzed dehydration/crosslinking yielding transient protective char) complemented by gas-phase inhibition (phosphorus-containing species), with the char ultimately oxidized in air at higher temperatures.

For broader comparability and design optimization, the recommended next steps include STA in N₂ (true char yield), TGA-FTIR/MS coupling for synchronized gas fingerprints, cone/microscale calorimetry to benchmark HRR/THR, and exploration of P–N synergy and coating uniformity/durability.