Introduction

The interest in nitroaromatic compounds over the years as important structural elements for the synthesis of industrial chemicals, dyes, pigments or herbicides is well recognized [1,2,3,4]. These compounds, despite the potential toxicity issues [5, 6], have been known as therapeutics for Parkinson’s disease, angina and insomnia [7,8,9], and recently their use in infectious diseases has been developed. Among these compounds, 4-nitrobenzoic acid has been listed as inhibitor agent for recognition of Mycobacterium tuberculosis complex and differentiation from non-tuberculous mycobacteria [10, 11]. Considering the present and future prospects of nitro drugs based on the unique properties of nitro group and its ability to form hydrogen bonds, there is a compelling need for complex researches in this controversial topic.

In pharmaceutical industry, the salts represent about a half of all active pharmaceutical ingredients on the market, bringing an improvement in drug’s physicochemical properties (e.g. melting point, chemical and physical stability, solubility, bioavailability and dissolution rate). In addition, selection of the appropriate anion/cation combination, keeping anion constant and changing the counterion or introduction of specific functional groups on the cation or anion can affect the physicochemical and biological properties of salt. An increasing number of 4-nitrobenzoate multicomponent complexes formed with organic cations such as aliphatic amines [12], guanidine [13, 14], 8-hydroxyquinoline [15], amino acids [16, 17], piperazine [18], pyridine [19], imidazole [20] are developed and studied recently in the literature as promising NLO materials or active pharmaceutical ingredients.

In the last years, searching for new compounds with potent biological activity and low toxicity seems to be a new trend in biosciences. Our previous studies highlighted the alkanolamines as promising molecules used as salt-forming agents and revealed the structural complexity and low toxicity of the system represented by alkanolamine salts with different aromatic carboxylic acids. These salts have provided interesting supramolecular networks with different topologies according to structural variation of counterions [21, 22], lower toxicity comparatively with aliphatic amines and heterocyclic amine salts [23] and biological activity in plants [24,25,26].

In this work we have focused on investigation of thermal behaviour and kinetics by thermal analysis techniques (DSC and TG/DTA) of alkanolammonium salts of 4-nitrobenzoate in order to highlight the nature of their solid state. Both counterions have active pharmaceutical properties and are economically, commercially available, which makes them promising for the design of new dual active compounds.

Materials and methods

All reagents, 4-nitrobenzoic acid—4-NBA, and the bases (ethanolamine—MEA, diethanolamine—DEA, triethanolamine—TEA) were purchased in analytical purity grade from Fluka AG (Buchs SG) and used as received.

The salts were prepared in a 1:1 molar ratio, dissolving the acid and the appropriate alkanolamine in acetone and then mixing the two solutions. The reaction was completed after 2 h of continuous stirring at room temperature, the salts precipitating in a crystalline state according to their solubility in acetone. The compounds were collected by filtration, washed with acetone and then dried in a vacuum for 3 h. The salts were recrystallized from acetone or ethanol. The suitable crystals for single-crystal X-ray diffraction analysis were obtained by slow evaporation at room temperature.

The purity of all studied salts was established via an UV spectrophotometric method, which consisted in determining the specific extinction (absorbance at λmax of a solution 1 gl−1 in a cell of 1 cm width) of the free acid (εa) and the corresponding salt (εs), by measuring the UV absorbance in NaOH 0.1 M at λmax.

$$ {\text{Purity}} = \frac{{\overline{\varepsilon }_{\text{s}} \cdot M_{\text{s}} }}{{\overline{\varepsilon }_{\text{a}} \cdot M_{\text{a}} }} \cdot 100 $$
(1)

where Ms and Ma are the molecular weights for the salt and free acid, respectively.

FTIR spectra were performed using PerkinElmer Spectrum 100 FTIR spectrometer with UATR, in the range 4000–650 cm−1 (see Fig. 1). Single-crystal X-ray diffraction analyses for all studied compounds were done at room temperature on a Xcalibur “Oxford Diffraction” diffractometer equipped with CCD area detector and a graphite monochromator utilizing MoKα radiation. Thermal behaviour was determined using a Diamond TG/DTA PerkinElmer in dynamic air atmosphere (synthetic air 5.0 Linde Gas with flow 100 mL min−1). The experiments were carried out from 25 °C up to 500 °C, at heating rates β = 5, 7, 10, 12 and 15 °C min−1, using open Al crucible (see Fig. 2). The phase transitions of the microcrystalline solids were observed using an Olympus BX53M polarizing microscope (POM) equipped with Linkam hot stage. Images of the various crystalline phases were recorded with an Olympus UC90 camera, using a filter with λ = 530 nm. Differential scanning calorimetry study was performed with one PerkinElmer Diamond DSC using aluminium closed pans with heating and cooling rate of 10 °C min−1 in nitrogen 20 mL min−1 (5.0 Linde Gas).

Fig. 1
figure 1

FTIR-UATR of 4-nitrobenzoic acid salts prepared with ethanolamine (A), diethanolamine (B) and triethanolamine (C). A—C9H12N2O5 (228.20); FTIR-UATR (1950νmax/cm−1): 2930, 2689 (–NH +3 ), 1646, 1612, 1556, 1385 (–COO−), 1102, 1068, 1021, 1009 (C–OH); B—C11H16N2O6 (272.26); FTIR-UATR (1942 νmax/cm−1): 3031, 2738 (–NH +2 ), 1615, 1567, 1377 (–COO−), 1098, 1066, 1035, 1007 (C–OH); C—C13H20N2O7 (316.31); FTIR-UATR (1980 νmax/cm−1): 2722 (-NH+), 1615, 1560, 1548, 1376 (–COO−), 1095, 1071, 1037, 1008 (C–OH)

Full size image
Fig. 2
figure 2

TG/DTG/feat flow (in air) of 4-nitrobenzoic acid salts prepared with ethanolamine (a), diethanolamine (b) and triethanolamine (c) (in air at heating rates β = 10 °C min−1)

Full size image

Results and discussion

Synthesis and physicochemical characterization

A series of alkanolammonium salts of 4-nitrobenzoic acid were prepared according to “rule of three”, via proton exchange reaction, each base having a sufficiently high pKa to react with 4-nitrobenzoic acid (Table 1). In this principle, the degree of ionization is considered to be a critical parameter, the ΔpKa (pKa of base − pKa of acid) value being important for predicting salt or co-crystal formation. ΔpKa value greater than 3 leads to salt formation, while ΔpKa less than 0 lead to co-crystal formation [27].

Table 1 Tabulated ΔpKa of alkanolammonium salt/protic ionic liquid of 4-nitrobenzoic acid
Full size table

The reaction has occurred rapidly, with high yields (90–92%). The compounds are water soluble and with well-defined melting points not exceeding ca. 0.5 °C, which is an indicative of their high purity. All synthesized alkanolamine salts have showed lower melting points compared with 4-NBA. The values of purity determined by spectrophotometric method were between 99.12 and 99.36% (Table 2). UV spectroscopic data showed similar λmax values for salt and acid (4-NBA λmax = 273.6 nm) that confirm the existence of the same anion in both compounds.

Table 2 The purity values of studied compounds
Full size table

The evidence of proton transfer and molecular interactions between alkanolammonium cation and 4-nitrobenzoate anion is illustrated in the FTIR-UATR spectra (Fig. 1). The formation of salts was identified by: (a) the presence of asymmetric and symmetric ν(COO−) vibrations occurring at 1650–1540 and 1450–1360 cm−1, (b) N–H vibrations attributable to –NH +3 , –NH +2 and –NH+, stretching of solid ammonium salts νNH3+= 3200–2700 cm−1, νNH2+= 3100–2700 cm−1, νNH+= 2800–2400 cm−1 and (c) the absence of stretching band νC=O = 1710–1680 cm−1, characteristic for the carboxyl group [29, 30]. The absorption bands reflected in FTIR-UATR spectra confirmed the conversion of carboxyl group of 4-NBA and amino group of alkanolamine into carboxylate and ammonium groups, indicating the complete proton transfer in studied compounds.

The crystallographic data of A–C compounds have revealed that the salt unit serves as building block for supramolecular architecture. The anions and cations are self-assembled via ionic N–H···O and normal O–H···O hydrogen bonds, compound A showing stronger intermolecular interactions [21], which leads to a higher melting point compared to compounds B and C. Therefore, the force of attraction between the molecules affects the melting point of a compound.

Thermogravimetric analysis of compounds A–C

The TG/DTG/DTA curves of studied compounds recorded in air at a heating rate of 10 °C min−1, and thermoanalytical data are presented in Fig. 2 and Tables 3 and 4, respectively. The thermoanalytical data obtained at 5 heating rates, β = 5, 7, 10, 12 and 15 °C min−1 were used for providing the kinetic parameters of the degradation process (FR, FWO and modified NPK kinetic methods).

Table 3 TG data of compounds A–C
Full size table
Table 4 Heat flow data of compounds A–C
Full size table

In the thermogravimetric study, similar behaviour for all samples was found. Samples B and C show one highly exothermic decomposition stage in the range of 150–300 °C.

Kinetic study

The kinetic analysis was performed using the TG data in air for the substance’s decomposition at five heating rates: 5, 7, 10, 12 and 15 °C min−1. To perform the kinetic analysis of the TG experimental data, three isoconversional methods were used, a differential one and two integral ones, respectively; fourth method elaborated by Sempere et al. [31, 32] and modified and developed by Vlase et al. [33, 34], the nonparametric kinetics method (NPK), in order to obtain kinetic parameters.

It is generally accepted that the reaction rate can be expressed as a product of two separable functions, k(T) dependent only on the reaction temperature T and f(α) dependent on the reaction degree α, i.e.:

$$ {\text{d}}\alpha /{\text{d}}t = k\left( T \right) \cdot f\left( \alpha \right) $$
(2)

where t is the reaction time. Under non-isothermal condition usually the reaction temperature depends linearly on time, T = Ti + βt, where β is the constant heating rate, so Eq. (1) becomes:

$$ \beta \cdot{\text{d}}\alpha /{\text{d}}T = k\left( T \right) \cdot f\left( \alpha \right) $$
(3)

Equation (3) is the starting point for all the data processing methods used in kinetic analysis. The TG/DTG curves furnish directly the dα/dT, α and T for different conversion degrees determined.

We used four different kinetic methods:

  • Differential isoconversional by Friedman [35] (FR)

    $$ ln\left( {\beta \cdot d\alpha /dT} \right)_{\alpha } = ln\left[ {A \cdot f\left( \alpha \right)} \right]_{\alpha } {-}E/RT $$
    (4)

  • Integral isoconversional by Flynn–Wall [36] and Ozawa [37] (FWO)

    $$ \ln \beta = \ln A/\left[ {R \cdot g\left( \alpha \right)} \right] \, {-} \, 5.331 \, {-} \, 1.052 \cdot E/R \cdot T $$
    (5)

    where g(α) = \( \int\limits_{0}^{\alpha } {} \)[dα/f(α)] is the integral form of the conversion function.

  • The Kissinger–Akahira–Sunose [38, 39] method (KAS) with Eq (6):

    $$ \ln \left( {\beta /T^{2} } \right) = \ln \left[ {A \cdot R/E \cdot g\left( \alpha \right)} \right] \, {-}E/R \cdot T $$
    (6)

By these three methods, the plotting of the left member of Eqs. (4), (5) and (6) versus 1/T, the value of the activation energy will be obtained from the line slope. The f(α) remain implicit, so these methods are “model free”.

By plotting the values of the activation energy E versus the conversion degree, the diagrams in Fig. 3 were obtained. Variations of E versus α exceed 10%, and these variations are non-monotonous. Therefore, a mean value of the activation energy is reasonable only for comparison (see Table 6).

Fig. 3
figure 3

Activation energy with errors bar versus conversion degree, determined by three different isoconversional methods for compounds A–C

Full size image

According to Eq. (3), the experimental data are plotted in a 3D coordinate system (dα/dT, T, α) and then interpolated for obtaining the reaction rate surface. This surface is partitioned as a i × j matrix M.

$$ {\mathbf{M}} = \left\{ {m_{{\text{ij}}} } \right\} = \left\{ {f\left( {T_{{\text{i}}} } \right)^{.} g\left( {\alpha_{{\text{j}}} } \right)} \right\} $$
(7)

The NPK method uses the singular value decomposition (SVD) algorithm to decompose matrix M into two vectors [49]. The matrix M is decomposed as follows:

$$ {\mathbf{M}} = {\mathbf{U}}\left( {diag\;{\mathbf{s}}} \right){\mathbf{V}}^{{\mathbf{T}}} $$
(8)

and the first columns of U and V matrix, respectively, u1 and v1 are analysed for determining the kinetic model, i.e.:

$$ {\mathbf{u}}_{1} = g\left( \alpha \right) $$
(9)

and a temperature dependence, i.e.:

$$ {\mathbf{v}}_{1} = f\left( T \right) $$
(10)

For the temperature dependence, the Arrhenius equation is selected, whereas for the kinetic model, we suggest the use of Śestak–Berggren equation [50]:

$$ g\left( \alpha \right) = \alpha^{{\text{m}}} \left( {1 - \alpha } \right)^{{\text{n}}} $$
(11)

where αm describes the influence of physical phenomenon related with the presence of reaction product and (1 − α)n describes the chemical phenomenon which involves the remainder reactant (1 − α).

If the decomposition process is a result of two or more simultaneous steps, it means that r = Σ ri and consequently the matrix M becomes:

$$ {\mathbf{M}} = \, \varSigma {\mathbf{M}}_{\text{i}} $$
(12)

The contribution of each step to the observed process is expressed by its variance, λ, so that Σλi = 100%. By the data processing with the NPK method, kinetic parameters were obtained (Table 5).

Table 5 Mean activation energy according to modified NPK method
Full size table

Activation energy value, presented in Table 6, indicates a reasonable agreement between the E values obtained by four different kinetic methods. The advantages of the NPK methods are: no supplementary approximations necessary for solving the temperature, and the respective conversion integrals and the kinetic description were complete.

Table 6 Mean activation energy (\( \overline{E} \) /kJ mol−1) determined by the four different kinetic analysis methods
Full size table

Calorimetric analysis of compounds A–C

DSC curves and melting characteristics of A–C compounds under N2 atmosphere are illustrated in Fig. 4 and Table 7. DSC curves of compounds A are characterized by the presence of one endotherm peak associated with melting with decomposition and the absence of solvate forms (Fig. 4), while for compounds B and C, a phase transition crystal–crystal was observed before the endotherm peak associated with melting. The transition between different phases was examined under Olympus BX53M polarizing microscope (POM) equipped with Linkam hot stage (Fig. 5).

Fig. 4
figure 4

DSC curves showing different situations at heating solid line, respectively, at cooling dashed line for A–C alkanolammonium salts

Full size image
Table 7 DSC characteristics of compounds A–C
Full size table
Fig. 5
figure 5

Optical micrographs of the crystalline solids B a and C b showing the transitions between different phases taken with filter λ = 530 nm

Full size image

Conclusions

New alkanolammonium salts based on pharmaceutical acid 4-NBA have been synthesised, and their thermal behaviour were compared. The studied compounds display various phase transitions in the solid state, which have been examined by thermal analysis, differential scanning calorimetry and hot-stage microscopy. Since in this study the 4-nitrobenzoate anion is a constant, differences of the solid state can be attributed to the alkyl and hydroxyl groups of the cation. When the cation possesses only one hydroxyl group (4-NBA MEA), the salt shows one endotherm peak associated with melting and decomposition. Substitution of H atoms linked to N atom in alkanolamines, with hydroxyl and ethyl groups, leads to different hydrogen-bonding network in final compounds and exhibits a phase transition crystal–crystal below the melting point. Compounds B and C show low melting points near 100 °C, compared to compound A which melts at 175 °C. The activation energies of the decomposition reactions were determined using four different methods, the values obtained by KAS and NPK being in good agreement. Kinetic results showed differences in activation energy between the studied compounds which increased in the order of C > A > B.