Thermal Behavior and Infrared Absorbance Bands of Citric Acid

Abstract

Citric acid is widely used in the Food and Pharmaceutical Industry. Various issues regarding its thermal behavior and infrared spectrum require clarification. Here, we studied citric acid monohydrate (raw, heated, freeze-dried and recrystallized from D₂O) via Differential Scanning Calorimetry, Thermogravimetric Analysis, Infrared Spectroscopy, and antioxidant capacity assay. Also, we used ab initio Density Functional Theory calculations for further supporting the interpretations of the experimental results. Citric acid monohydrate exhibits desolvation inability and upon heating does not dehydrate but esterifies. Nor by freeze drying can it be dehydrated. The heated sample is not anhydrous, it exhibits melting inability, and any fluidization occurs simultaneously with decomposition. In other words, the interpretations regarding the two endothermic peaks in the DSC curve of citric acid that have been attributed to water evaporation and melting are not correct. The increase in the molecular weight due to esterification is most likely responsible for the increased antioxidant/chelation capacity of the heated sample. We concluded that what we call citric acid monohydrate and anhydrous do not exist in a pure form (in the solid state) and actually are mixtures of different compositions of citric acid, water and a citric acid oligomer that is produced through esterification. The esterification reaction seems to be able to proceed easily under mild heating or even at room temperature. The presence of the ester oligomer and water affect the infrared spectrum of citric acid monohydrate and anhydrous and is responsible for the existence of multiple peaks in the C=O stretching region, which partially overlaps with the water H-O-H bending vibration. The insights presented in this work could be useful for optimizing the design, performance and quality of food and drug products in which citric acid is used.

1. Introduction

Thermal behavior, e.g., the melting point and the thermal stability of materials, is of great importance and is involved in many applications, including material processing, food and drug preparation, etc. Recently, it has been reported that various substances, such as polymers [1], flavonoids [2] and gallic acid [3], upon heating, do not exhibit the typical thermophysical transition of melting, and instead, they decompose. In many cases, the decomposition and “melting” (material softening or fluidization) overlap and occur simultaneously. The term “thermochemical transition” was proposed for this effect [1]. In some cases, e.g., gallic acid [3], in addition to the solid–liquid thermochemical transition at temperatures above 200 °C, a solid–solid thermochemical transition was observed to occur at temperatures below 100 °C. This behavior was explained based on the comparison of the Gibbs free energies of decomposition and melting and was put in the broader frame of “melting inability” [4].

In addition to the term melting inability, the term “desolvation inability” was introduced in order to describe the occurrence of decomposition of a solute prior to being able to accomplish the solvent’s full removal upon the heating of a solvate/hydrate [5]. An explanation, similar to that for melting inability, has been provided for the desolvation inability; that is, the Gibbs free energy of decomposition is lower than the one of desolvation/dehydration [5]. The desolvation inability is the actual cause of the above-mentioned solid–solid thermochemical transition of gallic acid below 100 °C. In substances like flavonoids and gallic acid, which exhibit antioxidant properties, it would be interesting to examine if and how the antioxidant activity is influenced by the minor decomposition that occurs during the solid–solid thermochemical transition. For substances with melting and/or desolvation inability, a confusion regarding their thermal behavior has been recognized. For example, the solid–solid thermochemical transition of gallic acid has been interpreted as water evaporation or solid–solid crystallographic (thermophysical) transformation [3]. Also, some confusion was reported for its infrared (IR) absorbance bands [3]. Of course, the knowledge of the IR bands for a given a substance is very useful on many occasions.

The primary decomposition pathway for gallic acid was found to be an esterification reaction. Also, it has been reported that the thermal behavior of cellulose esters is governed to great extent by the esterification reaction between the polymer and its acid impurities, which can be produced by the hydrolysis of the polymer [1]. The easiness of the occurrence of the esterification reaction was reported to be related to the low value of the enthalpy of esterification [1]. As mentioned above, a thermodynamic explanation has been reported for the melting inability, and also this behavior has been predicted for various organic substances including various hydroxyl-acids, e.g., citric acid [4]. Citric acid is a naturally occurring hydroxy-acid that is widely used as an antioxidant, a preservative, a stabilizer, etc., in the Food [6,7] and Pharmaceutical [8,9] industry. Citric acid is believed to exist in anhydrous and monohydrate forms. Confusion regarding the thermal behavior of citric acid can be recognized in the literature. Although it has been reported that the melting and decomposition of anhydrous citric acid occur at very close, practically overlapping, temperature ranges [10], in various cases [11,12,13,14], including the NIST Chemistry Webbook [12] and the Handbook of Chemistry and Physics [11], typical melting points are reported in the range of 150–160 °C. In the Handbook of Chemistry and Physics, a melting point of 135 °C is reported for citric acid monohydrate and a melting point of 153 °C is reported for anhydrous citric acid [11]. Also, no information is given for the boiling point of the monohydrate form, while for the anhydrous form, is it mentioned that it decomposes and does not boil [11]. The monohydrate form, in addition to the melting peak, exhibits an endothermic peak in its Differential Scanning Calorimetry (DSC) curve below 100 °C [15,16]. Anhydrous citric acid is believed to be able to be prepared by heating, e.g., drying at 100 °C to remove absorbed moisture [10], while citric acid monohydrate is believed to dehydrate, and it has been reported that the onset temperature of its dehydration is 71 °C [16]. Of course, if the monohydrate form is able to dehydrate, then it should not exhibit any melting point, and there should not be any difference in the melting point of the anhydrous and hydrate form. However, as mentioned above, in the Handbook of Chemistry and Physics quite different values are given for the melting points.

In addition to the thermal behavior, an interesting aspect can be recognized in the IR spectrum of citric acid and citric acid monohydrate. More precisely, in the C=O stretching region (~1650–1800 cm⁻¹), citric acid monohydrate exhibits a triple peak [17], while citric acid exhibits a double peak [12]. However, it is widely known [18] that the C=O stretching IR bands are very characteristic depending on the origin of the C=O group (that is, the C=O stretching of ketones, acids, esters, etc., appears at distinct and characteristic wavenumbers). In both citric acid and citric acid monohydrate, only one type (acid) of the C=O group is supposed to exist; thus, only one respective peak should be present in the IR spectrum. However, as mentioned above, this is not the case. In addition, in the IR spectrum of citric acid, one sharp peak in the O-H stretching region (~3100–3600 cm⁻¹) is present [12]. Such sharp peaks are not typical of the O-H bands of organic molecules. In addition, this sharp peak seems to be the same peak that was observed in gallic acid and was assigned to water O-H stretching. In a few words, there are various aspects that require clarification regarding the thermal behavior and the IR absorbance bands of citric acid and its monohydrate form. The scope of this work is to provide insights regarding these issues. More precisely, we provide a comprehensive analysis of the thermal behavior and the infrared spectrum of citric acid monohydrate and citric acid. We found that citric acid monohydrate cannot be dehydrated without being decomposed, and neither the hydrate nor the non-hydrate form exhibits a melting point. It is shown that citric acid does not exists in pure form but only as a mixture of citric acid, water and a citric acid oligomer that is produced through esterification. These strongly affect its IR spectrum. It is also shown that the partial alteration of the chemical structure of citric acid can occur easily by mild thermal treatment or even at room temperature. The alteration of the chemical structure affects its antioxidant activity.

2. Materials and Methods

Citric acid monohydrate (Food Grade) was purchased from a local market (Volos, Greece). KBr (>99.5%) was purchased from Chem-Lab (Zedelgem, Belgium). D₂O (99.9%) was purchased from Deutero GmbH (Kastellaun, Germany). Trolox (6-hydroxy-2,5,7,8-tetramethylchroman-2-carboxylic acid) was purchased from Sigma (St. Louis, MO, USA). The KBr was dried at 140 °C for 2 h prior to use, while all other chemicals were used as received.

Thermogravimetric analysis (TGA) measurements were carried out using a Shimadzu TGA-50 thermogravimetric analyzer (Shimadzu, Tokyo, Japan). Differential Scanning Calorimetry (DSC) measurements were performed using a Shimadzu DSC-50 calorimeter (Shimadzu, Tokyo, Japan). Fourier Transform Infrared (FTIR) measurements were performed using a Thermo FTIR spectrometer (Thermo Fisher Scientific, Waltham, MA, USA). Antioxidant activity was evaluated spectrophotometrically. A Christ freeze dryer (model Gamma 1-20, Martin Christ Gefriertrocknungsanlagen, Osterode am Harz, Germany) was used for the preparation of the freeze-dried sample.

Four samples were examined, namely, (a) raw citric acid monohydrate, (b) citric acid monohydrate heated at 100 °C, (c) citric acid monohydrate freeze-dried and (d) citric acid recrystallized from a D₂O solution. The recrystallized sample was prepared by solution casting. The solution was left at room temperature for 3 days for complete solvent evaporation. For the freeze-drying process, initially the sample was frozen at −20 °C under atmospheric pressure. Then, vacuum was applied, and simultaneously the plate in which the sample was placed was heated in order to provide the latent heat of sublimation. During freeze drying, the temperature of the drier was stabilized at −12 °C.

The raw citric acid sample was examined with DSC, TGA and FTIR. All other samples were examined with FTIR. Aqueous solutions of the raw and the heated samples were examined for their antioxidant activity. The Ferric Reducing/Antioxidant Power (FRAP) assay was performed using the modified method by Pulido et al. [19]. In brief, 100 μL of the raw and heated samples (0.1 g/mL) was mixed with 2.9 mL of freshly prepared FRAP reagent. Samples were mixed and incubated at 37 °C for 10 min, and the absorbance of each sample was measured at 593 nm. Trolox was used as a standard for the calibration curve. The results were expressed in μΜ of Trolox equivalents. The DSC and TGA measurements were performed with a heating rate of 10 K/min from room temperature up to 165 °C under nitrogen atmosphere (20 mL/min flow). The FTIR measurements were carried out with a resolution of 2 cm⁻¹ by performing 64 scans in absorbance mode. Baseline correction and normalization at a scale of 0–1 was applied in all the FTIR spectra. In order to facilitate the comparison of the spectra of the samples, the approach of spectra subtraction was adopted.

3. Ab Initio DFT Calculations

One of the IUPAC names of citric acid is 3-carboxy-3-hydroxypentane-1,5-dioic acid [12]. We used this name as the basis in order to give code names to the possible decomposition products of citric acid. According to this name, it can be realized that citric acid has three COOH groups at positions 1, 3 and 5 and one OH group at position 3. Since there are three COOH groups and one OH group in citric acid, there are two possible different esters that can be produced (by considering the COOH groups at positions 1 and 5 equivalent). In addition, the esterification reaction can take place either inter-molecularly (that is, between two different citric acid molecules) to give a dimer or intra-molecularly (that is, within the same molecule) to give a cyclic substance. Also, a dehydration reaction between the COOH groups is possible (again both inter- and intra-molecularly) to give anhydride. In total, there are nine possible reactions with nine decomposition products along with water as the second decomposition product. In order to avoid complex and large names and in order to facilitate the presentation and discussion of the results, we give a code name to these products. The code name is composed of one letter, two numbers and the term “dim”, e.g., E13dim, A15. The letter may be A or E, corresponding to an anhydride or ester product, respectively. In the anhydrides, the numbers correspond to the positions of the COOH groups that take part in the reaction. In esters, the first number corresponds to the position of the COOH group, while the second number is 3, corresponding to the position of the OH group. The term “dim” refers to dimers and thus is used for the products produced by inter-molecular reactions. For example, the name A13dim corresponds to the anhydride that is formed through the reaction of the COOH group at position 1 of one citric acid molecule with the COOH group at position 3 of another molecule. The name E13 is the cyclic ester produced by the reaction of the COOH group at position 1 with the OH group at position 3 within the same citric acid molecule. As mentioned above, the COOH at position 1 and 5 are considered equivalent, and thus, the E53dim is the same as the E13dim.

In this work, we used DFT to find the Gibbs free energy of the above-mentioned nine reactions in order to determine which of the nine products is the most probable product. According to the literature, the PBE0 functional with a TZ basis size is accurate for structure and frequency optimization and of intermediate accuracy, as is the other functional for estimating energies of reactions [20]. Also, it is recommended to use a dispersion correction in the calculations [20]. Thus, we used the PBE0 [21,22] functional with the def2-TZVP [23] basis set along with the def2/J [24] auxiliary basis and the D4 [25,26] dispersion correction.

The molecules were drawn in Avogadro software (Windows version 1.1.0) [27,28], and an initial (rough) geometry optimization was also performed with Avogadro software. These atom coordinates were used as input for the geometry optimization with ORCA software [29,30,31]. The frequency optimization was also performed with ORCA software [29,30,31]. For the calculations, ORCA uses the libint2 library for the computation of 2-el integrals [32]. Rotational and vibrational entropy was computed in ORCA according to Herzberg [33] and the QRRHO of Grimme [34], respectively. We used version 5.0.2 of ORCA, which was built with support for libXC version 5.1.0 [35].

4. Results and Discussion

4.1. Raw and Heated at 100 °C Citric Acid Monohydrate Samples

4.1.1. Thermal Behavior and IR Spectra

In Figure 1, the DSC and TGA curves of citric acid monohydrate are presented in the range of 40–165 °C. As can be seen, in this temperature range, citric acid monohydrate exhibits two endothermic effects (DSC curve), and both of them overlap with considerable mass loss (TGA curve). The first thermal effect is centered at 61 °C, while the second effect is located at 150 °C.

According to the established interpretations, we could explain the above results as follows: citric acid monohydrate dehydrates, the mass loss up to 90 °C is water evaporation, and then the anhydrous citric acid melts at 150 °C [11,12]. However, we dispute that such an interpretation is correct. Although the temperature of 150 °C is very close to the literature values [11,12] of the melting point of citric acid, it is clear that this effect overlaps with mass loss. In Figure 2, photographs are presented that show inside the TGA pan of the raw citric acid monohydrate sample and the same sample after being heated up to 165 °C (the temperature of 165 °C was selected since at this temperature the thermal effect is completed, as can be seen in the DSC curve in Figure 1).

It is clear from the photos of Figure 2 that the sample undergoes severe decomposition by heating up to 165 °C. As mentioned above, the center of the second peak in the DSC curve is located at 150 °C, and the end of this thermal effect is located in the range of 160–165 °C. If this thermal effect was melting, then no sign of decomposition should be observed after heating the sample at this temperature range. Thus, it is clear that the thermal effect that is centered at 150 °C and ends at 165 °C cannot be melting. Citric acid exhibits melting inability. Actually, it is not citric acid that exhibits melting inability but a mixture of citric acid, water and the ester of citric acid, as we will discuss next. More precisely, we will now argue that the broad endothermic peak at 61 °C does not arise solely from the heat, which is required for water evaporation, but there is a contribution from the heat, which is required for the esterification reaction that occurs upon the heating of the sample. In order to support this claim, we will focus on the thermal effect below 100 °C, which is centered at 61 °C. In order to study this effect, a citric acid monohydrate sample was heated up to 100 °C. The temperature of 100 °C was selected since at this temperature, the mass loss of the first stage has been completed (see TGA curve in Figure 1) and the thermal effect at 61 °C seems to be completed (see DSC curve in Figure 1). The FTIR spectra of the raw and the heated at 100 °C samples along with their subtraction spectrum are presented in Figure 3.

As can be seen in Figure 3a, the IR spectrum of the two samples is quite different. Obvious alterations can be seen in the region of O-H stretching (3100–3700 cm⁻¹). Also, in Figure 3b, the same spectra are presented around the region of C=O stretching in order to show the double peak of the heated sample and the triple peak of the monohydrate sample, which were mentioned in the Introduction. The presence of multiple distinct C=O stretching vibrations indicates the existence of multiple types of C=O. The wavenumbers of the triple peak are 1690, 1727 and 1753 cm⁻¹. The peak at the higher wavenumber at 1753 cm⁻¹ cannot be assigned to acid groups since it is characteristic of aliphatic ester C=O or anhydride C=O, which typically appear at even higher wavenumbers, e.g., 1740–1800 cm⁻¹ [18]. It is known that the C=O stretching vibration of carboxylic acids appears at quite lower wavenumbers, e.g., 1700 cm⁻¹ [18]. Thus, the peak at 1753 cm⁻¹ cannot be assigned to acid C=O and points out the existence of ester C=O. In addition, the water bending vibration occurs at around 1650 cm⁻¹, but no such peak is present in the spectrum of citric acid monohydrate. In the case of gallic acid, shifting of the water bending peak was observed, which depended on if the water molecule interacted with the COOH or OH group [3]. Thus, it seems likely that the water bending band in the spectrum of the citric acid monohydrate is shifted at higher wavenumbers and overlaps with the C=O stretching of acid groups. Thus, the triple peak of the citric acid monohydrate in the C=O region contains contributions from acid C=O stretching, ester C=O stretching and water H-O-H bending. This means that the citric acid monohydrate does not exist in the solid state, it partially decomposes through esterification, and the substance we name citric acid monohydrate is actually a mixture of citric acid, water and citric acid ester, most likely in equilibrium. In the region of O-H stretching (3100–3700 cm⁻¹), the characteristic bands of water O-H stretching at 3490 and 3280 cm⁻¹ [18] cannot be distinguished in the IR spectrum of citric acid monohydrate since they overlap with the O-H stretching of the OH group at higher wavenumbers and the O-H stretching of COOH groups, which is known to occur at lower wavenumbers, e.g., in the region 2500–3300 cm⁻¹ [18].

In the heated sample, two sharp peaks at 3494 and 3290 cm⁻¹ are present. The two peaks are very close to the literature values of water at 3490 and 3280 cm⁻¹ [18]. These two peaks were also present in the IR spectrum of gallic acid and were found to be related to water O-H stretching [3]. The IR spectrum of gallic acid is presented in Figure 3c, in which the resemblance with the spectrum of the heated citric acid can be seen in the region 3000–3600 cm⁻¹. The above-mentioned peaks and especially the sharp peak at 3498 cm⁻¹ are clearly visible in both spectra. In the subtracted spectrum (Figure 3a,b), it is clear that these are two positive peaks inside the overall negative O-H stretching peak. This clearly points out that the overall OH groups have decreased, with a simultaneous increase in the water content. The esterification reaction could explain this alteration in the spectrum of the heated sample, since it should result in a decrease in O-H stretching of the COOH and OH groups and cause an increase in the water bands. Of course, in such a case, the water bending band should also be positive, and the ester C=O band should also increase. Indeed, in the subtracted spectrum, it can be seen that two positive peaks exist at 1745 cm⁻¹ (assigned to ester C=O) and at 1705 cm⁻¹ (assigned to water H-O-H bending). The negative peak at this region can be related to the reduction of the acid C=O groups due to the esterification reaction. Based on the above, we propose the band assignments that are presented in

Table 1. Proposed IR band assignments for citric acid monohydrate in the region 1650–4000 cm⁻¹.

Wavenumber, cm−1	Citric Acid Monohydrate
3340–3700	Overlapping of O-H stretching vibration of water and alcohol OH groups; the contribution of water is around 3490 cm−1
3000–3340	O-H stretching vibration of acid OH groups with water contribution at around 3280 cm−1
1745–1800	C=O stretching of ester carbonyl
1704–1745	C=O stretching of acid carbonyl overlapping with ester C=O at higher wavenumbers
1650–1704	Overlapping of H-O-H bending with C=O stretching of acid carbonyl

and

Table 2. Proposed IR band assignments for citric acid anhydrous (citric acid monohydrate heated at 100 °C) in the region 1650–4000 cm⁻¹.

Wavenumber, cm−1	Citric Acid
3498	O-H stretching vibration of water
3445	O-H stretching vibration of alcohol OH
3291	Mostly O-H stretching vibration of water overlapping with acid OH groups
3000–3340	O-H stretching vibration of acid OH groups
1727–1800	C=O stretching of ester carbonyl
1650–1727	Overlapping of H-O-H bending with C=O stretching of acid carbonyl

for citric acid monohydrate and citric acid anhydrous (heated at 100 °C), respectively.

As a summary of this section, we can say that citric acid monohydrate undergoes esterification to some extent even at room temperature during its preparation/storage. Upon heating, it exhibits desolvation inability and does not dehydrate. Instead, by heating, the equilibrium between citric acid, its ester and water is disturbed and is shifted towards the direction of esterification, leading to the formation of more citric acid ester. This mixture (“anhydrous citric acid”) exhibits melting inability, and any fluidization occurs simultaneously with decomposition upon further heating.

4.1.2. Antioxidant Activity

As discussed in the previous section, the citric acid monohydrate sample undergoes a structure alteration through esterification at temperatures below 100 °C. Such temperatures may be used during food or drug processing where citric acid is used as an antioxidant, flavoring, a preserving agent, etc. Thus, we examined the antioxidant activity of aqueous solutions of the raw and the heated at 100 °C citric acid samples. The antioxidant capacity of the raw sample was found to be equal to 257 μM TE, while the heated sample exhibited a higher capacity of 312 μΜ TE. It should be stressed that citric acid does not have a direct antioxidant activity [36]; that is, it is not a primary antioxidant, but it can act as secondary antioxidant by inhibiting oxidation through chelation [36,37]. The antioxidant activity/chelation in this assay (which involves iron reduction) is related to the increased number of OH groups present [37], and since esterification leads to the reduction of such groups, the opposite result would be expected. The increase in the antioxidant capacity of the heated sample may be understood in terms of the increase in the molecular weight. More precisely, the esterification may take place between various citric acid molecules, and an oligomer (e.g., dimer, trimer, etc.) may be formed. Per mole of citric acid monohydrate, theoretically there are 4 moles of OH groups. Per mole of citric acid dimer, there are 6 moles of OH groups. This increase may explain the increase in the antioxidant activity. Also, it must be taken into account that inside water, the excess of water, according to Le Chatelier’s Principle, may favor the reverse reaction (hydrolysis), and more than six OH groups may be presented per mole of the dimer after hydrolysis. Finally, the different water content of the samples may also affect the observed alteration.

4.2. Recrystallized Sample from D₂O Solution

In Figure 4a, the IR spectra of the raw sample and the recrystallized citric acid sample from the D₂O solution are presented, along with their subtracted spectrum. No signs of the presence of D₂O can be detected (that is, the D-O stretching at 2540 and 2450 cm⁻¹ and the D-O-D bending at 1215 cm⁻¹ [18]). Interestingly, the spectrum of the recrystallized sample is pretty similar to the spectrum of the heated at 100 °C sample and similar conclusions are derived from the comparison of the spectra of the recrystallized and the raw sample. That is, during solvent evaporation at room temperature, it seems that esterification took place, leading to alterations in the spectrum similar to the ones caused by heating at 100 °C. In an excess of water, as mentioned above, the favoring of the hydrolysis direction is expected. However, as the evaporation of water proceeds and its concentration is decreased, the ongoing evaporation of water, according to Le Chatelier’s Principle, shifts the reaction towards the esterification direction. This could explain the coincidence of the spectra of the heated and recrystallized samples.

As mentioned in the previous section, if the esterification proceeds between different citric acid molecules (and not within the same molecule to give a cyclic substance), then an increase in the molecular weight is expected due to the formation of oligomers. Indeed, after recrystallization, as shown in Figure 4b, the sample was not in a powder form (behavior typical of low-molecular-weight substances) but in a film-like form (typical behavior of polymers). Of course, the film did not exhibit the mechanical coherence of typical macromolecules, but this poor ability for film formation clearly points out an increase in the molecular weight. This can be considered as additional evidence for the occurrence of inter-molecular esterification.

4.3. Freeze-Dried Sample

In Figure 5a, the IR spectra of the raw and the freeze-dried citric acid monohydrate sample are presented, along with their subtracted spectrum. As can be seen, the differences in the two spectra are less severe compared to the ones of the heated and the recrystallized samples. Since we have concluded that heating favors the esterification reaction, it would be expected that the decrease in temperature will favor the reverse reaction, and thus it is expected that esterification would proceed at a lower extent in the case of freeze drying. In addition, due to the low temperatures that are used during freeze drying, the reaction rate is expected to be quite lower. From the subtracted spectra, it is clear that only one positive peak exists at 1726 cm⁻¹. This peak could be attributed to a very small increase in the water bending band or/and the C=O ester stretching band. In such case, this means that a very minor decomposition has occurred during freeze drying. It is worth mentioning that in the case of gallic acid, clear signs of decomposition during freeze drying have been reported [5]. In any case, there is no evidence in the FTIR spectra to claim that water has been removed or even decreased after 21 h of freeze drying. This is another indication of the desolvation inability of citric acid monohydrate. For comparative analysis, in Figure 5b, the spectra of the raw, recrystallized and freeze-dried samples are presented.

4.4. DFT Calculations

In

Table 3. ΔG of the nine possible decomposition reactions of citric acid (formation of ester or anhydride) as calculated from DFT.

Product	ΔG of Reaction, kJ/mol
E33dim	+1.6
A13	+13.9
A15	+29.1
A13dim	+42.9
E13	+44.5
A11dim	+46.1
A33dim	+57.3
E13dim	+58.5
E33	+115.2

, the values of the Gibbs free energy, as calculated from DFT, of the nine possible reactions of citric acid esterification or anhydride formation are presented. As can be seen, seven out of the nine reactions have ΔG values in the order of magnitude of 10–100 kJ/mol, one reaction has a ΔG more than 100 kJ/mol, and one reaction has a very low (as an absolute value) ΔG value close to 1 kJ/mol. This reaction is the reaction that leads to the formation of a dimer through esterification between the OH group at position 3 of one molecule and the COOH group at position 3 of another citric acid molecule. Thus, the most probable product is the E33dim. This is in agreement with the interpretations of the experimental results for the formation of oligomers and the increase in molecular weight. Also, despite the fact that the ΔG values, as calculated from DFT, refer to gas-phase reactions, the very low value of ΔG implies and may explain the easiness of the occurrence of this reaction. This is also in agreement with the conclusions of the previous sections, where it was claimed that the esterification reaction takes place even at room temperature. In addition to the E33dim product, other probable products are mainly cyclic anhydrides, e.g., A13 and A15. These products may be formed during the decomposition at 150 °C or to a lower extent at lower temperatures.

5. Further Discussion

The results of this study showed that citric acid undergoes alterations of its chemical structure rather easily by a mild thermal treatment or even at room temperature. Since the esterification reaction is a two-direction reaction and can be reversed, the extent of esterification depends on temperature, available water, air humidity, etc. Thus, citric acid can undergo different alterations of its structure depending on its storage conditions, subjected treatment, etc. The provided insights could assist in improving the performance and quality of food or drug formulations that involve the use of citric acid. More precisely, based on the results of this study, it is clear that citric acid does not always exhibit the very same properties, which seem to depend on the subjected treatment as well as its storage conditions. For achieving the optimum design of a food or drug formulation, it is necessary to use the proper amounts of additives such as preservatives, antioxidants, etc.; that is, there should not be an amount of additive higher than the one needed, and on the other hand, the amount of the additive should be sufficient in order to be active and adequate to fulfill its purpose of use. The alteration of the chemical structure of citric acid obviously affects its properties either directly or indirectly. An example of direct influence, as shown in this study, is the alteration of its antioxidant activity, which can be related to the ability for chelation capacity. The alteration of the chemical structure due to various treatment or storage conditions affects the pH of citric acid and thus can also affect, indirectly, its preservative activity, which is related to its pH value. Similarly, citric acid is used to improve the solubility/release of drugs that are poorly soluble in water. The alteration of the structure and the reduction of OH and COOH groups can also affect the solubility and hydrophilicity of citric acid. All in all, different batches of citric acid may exhibit different chemical structures and thus different properties. Thus, such issues should be taken into account in order to achieve the optimum performance/quality of a product.

6. Conclusions

Citric acid does not seem to exist in pure form in the solid state either in the hydrate or anhydrous form. In the solid state, it exists as a mixture. More precisely, the multiple bands of the IR spectrum in the C=O stretching region point outs that what we call citric acid monohydrate is a mixture of citric acid, the citric acid oligomer produced through esterification and water. This complex hydrate exhibits desolvation inability, and upon heating it decomposes through esterification, leading to the formation of more ester product. Citric acid subjected to heating contains water and ester product. It exhibits melting inability (does not exhibit any melting point), and upon heating any fluidization occurs simultaneously with decomposition. In addition to the experimental evidence, ab initio DFT calculations indicate that the most probable product is the one produced through esterification between the OH group at position 3 of one molecule and the COOH group at position 3 of another citric acid molecule (the positions refer to the following IUPAC name of citric acid: 3-carboxy-3-hydroxypentane-1,5-dioic acid). Other probable decomposition products are cyclic anhydrides. The increase in the molecular weight during esterification along with the reversion of the reaction inside water may explain the increase in the antioxidant activity of the heated (“anhydrous”) sample over the one of the raw citric acid monohydrate sample. The monohydrate form cannot be dehydrated either by heating or by freeze drying.

In few words, citric acid undergoes alterations of its chemical structure by mild heating or even at room temperature. The primary reaction that is responsible for this alteration is an esterification reaction. The extent of esterification (which is a two-direction reaction and can be reversed to some extent) depends on temperature or an excess of products or reactants and thus depends on the subjected treatment and storage conditions. Thus, citric acid samples that have been subjected to different treatment or storage conditions are likely to exhibit different structures and thus different properties, e.g., different antioxidant activities. The provided insights could be useful for optimizing the design and performance of food and drug formulations that involve the use of citric acid.