Performance Evaluation of a Portable Laser-Induced Fluorescence Spectroscopy System for the Assessment of the Humification Degree of the Soil Organic Matter

Laser-induced fluorescence spectroscopy (LIFS) has shown advantages for analyses of whole soil, providing results about chemical structure of soil organic matter (SOM) without the need for a chemical fractionating process; thus, allowing direct analysis of soils. Moreover, another advantage is the potential for the development of portable systems, which could be usable in field analyses. The aim of this paper is to evaluate the performance of a portable LIFS system to assess the humification degree (HD) of SOM. A good correlation (R > 0.8) was obtained for results (HD) achieved using the proposed portable system and a conventional fluorescence spectrometric method for humic acid. In addition, the results obtained using the portable LIFS system statistically agreed with the ones achieved using a benchtop LIFS system already described in the literature for such application (R = 0.953). Thus, the portable LIFS system constitutes an affordable option for the in situ characterization of SOM.


Introduction
Nowadays, the concern about soil quality has increased considerably due to environmental problems related to land degradation and sustainable agricultural production.However, the definition of soil quality and the development of techniques for such evaluation is an important challenge for science.3][4] SOM comprises humic substances (HS) and non-humic substances.However, an important information, often disregarded when studying SOM, is the soil humification degree (HD).
8][19][20][21][22][23][24] For instance, EPR provides the semiquinone-type free radical concentration and 13 C NMR provides information of the degree of aromaticity. 8,25,26owever, the use of these techniques requires chemical extraction and fractionation of HS, which might modify fulvic acid, humic acid and humin content in respect to what is present of them in the original sample. 27,28Furthermore, sample preparation is labor-intensive, increasing analysis time and cost.
With regard to agricultural and environmental issues, it is extremely important to develop an analytical method able to provide a high analytical frequency (number of samples analyzed per unit of time) and able to meet the demand for information on HD and SOM.
This study makes a contribution of special interest for analytical chemistry because the portable laser-induced fluorescence spectroscopy system (LIFS) is a very promising technique to assess the HD of organic matter (OM) in whole soil samples and OM insoluble, as the humin.
Currently, in the literature, there is no technique able to determine the humification degree of the whole OM and the most of existing studies use HA, which represents only a fraction of the bulk SOM.Information based on humin, insoluble fraction of HS and that compose more than 80% of the humic material from SOM, is not reliable as measurements made using 13 C NMR and EPR suffer interference due to the presence of high levels of clay materials, iron oxide and other minerals that are paramagnetic, causing reduction in the relaxation times and degrading the resolution of the spectra. 13,26,29The advantage of LIFS is that it is a technique that enables the direct optical probing of substances in solid materials and can enable information on HD of the whole OM without the need for extraction and fractionation of HS from soil.
Several humification indexes based on fluorescence spectroscopy have been proposed. 21,22,24Zsolnay et al., 22 working with dissolved organic matter extracted from soil, proposed a humification index based on the ratio of the signals (integrated area) of the emission bands measured between 570 to 641 nm (A 4 ) and between 356 to 432 nm (A 1 ), using excitation as 240 nm.This index is referred as A 4 /A 1 .This definition is based on the fact that humification can be associated to an increasing of C/H ratio and a shift to longer fluorescence emission wavelengths. 8,30,31][35] Kalbitz et al. 21and Milori et al. 24 also defined humification indexes based on the spectral characteristics of humic acids in solubilized samples, they proposed another humification index denominated I 400 / I 360 , based on the ratio of fluorescence intensities measured respectively at 400 nm (I 400 ) and at 360 nm (I 360 ).Alternatively, it was used a similar index I 470 /I 360 with these intensities measures using fluorescence synchronous scanning in order to improve spectral selectivity.Milori et al. 24 presented the third humification index by measuring the fluorescence emission spectra of solubilized humic acids using excitation at 465 nm, index denominated A 465 .
LIFS has been used to assess the humification index of SOM in whole soil samples. 36The method is promising because it enables fast acquisition of results relying on the direct probing of the solid sample, allowing the assessment of OM associated with minerals in the soil matrix. 26The results are in agreement with the ones obtained using other spectroscopic approaches such as conventional fluorescence spectroscopy, EPR and 13 C NMR. 24,26,36 According to González-Pérez et al., 26 LIFS is a promising technique for OM studies with whole Oxisol samples.][38][39][40] The proposed LIFS method relies on the excitation of SOM using a laser in the near ultraviolet-blue region (350-480 nm), which coincides with the absorption profile of more rigid and complex molecules of HS, 24,41 such as aromatics and heterocyclic compounds.The aim of this study was to evaluate the performance of a lab-made portable LIFS system for the assessment of the humification degree of the SOM.The results of the portable LIFS system were compared with the values of humification degree of HA extracted from the same samples and determined by fluorescence spectroscopy methods. 21,22,24The performance of the portable system was also compared to a benchtop LIFS system similar to that described by Milori et al. 36

Experimental samples
Soil used in this study was the Typic Hapludults, cultivated with Tifton 85 Bermudagrass.The field experiment was located in the city of Lins, São Paulo.A set of soil samples with low carbon content was collected from six depths: 0-10 cm, 10-20 cm, 20-40 cm, 40-60 cm, 60-80 cm and 80-100 cm, being used a total of 36 samples (6 sets with 6 samples each).The samples were named A1 to A36.This set of samples was chosen due to its variation of humification degree of SOM.

Total organic carbon (TOC)
In the laboratory, the soil samples were sieved to remove roots and ground to obtain particles smaller than 0.15 mm.For this analysis, aliquots of 100 mg of each of the six samples of soil were used for all treatments.These aliquots were oxidized at 900 °C using a flow rate of 0.3 L min -1 oxygen.The determination of the percentage of Vol. 26, No. 4, 2015   TOC was performed in triplicate for samples of whole soil.The measurements were performed using a Total Organic Carbon Analyzer, TOC-V Model, Shimadzu Model, coupled to a Solid Sample Module, SSM-500A, Shimadzu.

Humification degree of humic acids assessed by conventional fluorescence spectroscopy
The HAs were extracted according to the procedure advised by the International Humic Substances Society (IHSS). 42he HAs were brought to a concentration of 20 mg L -1 and the pH was adjusted to 8.0 by diluting them in a solution of NaHCO 3 0.05 mol L -1 .This concentration of 20 mg L -1 was used to minimize post-filter effect and guarantee that analytical response is within its linear range.
Normal fluorescence spectra and synchronous fluorescence spectra (excitation band) were acquired using a Perkin Elmer Luminescence Spectrometer LS 50 B spectrofluorophotometer (right angle geometry, 1 cm × 1 cm quartz cell) using a 1 cm optical pathlength quartz cuvette, spectral band pass of 10 nm and scan velocity of 500 nm min -1 for both excitation and emission.
To obtain the A 4 /A 1 humification index, the emission spectra were scanned from 300 to 700 nm with excitation at 240 nm.The A 4 /A 1 ratio was calculated from the integrated area measured in the last quarter of the emission spectrum (570-641 nm) and the first quarter (356-432 nm). 22In this study, synchronous scan excitation spectra were acquired over a range of 220 to 650 nm using a Δλ of 55 nm and the humification index was calculated by the ratio of the fluorescence intensities measured at 468 and 374 nm (I 468 /I 374 ). 21To obtain the A 465 humification index, the emission spectra were obtained over a range of 480 to 800 nm using an excitation wavelength of 465 nm.In this case, the humification degree was calculated using the total area of emission (A 465 ). 24mification degree of SOM assessed using LIFS After the sample collection was performed, the samples were dried at room temperature, passed through a 100-mesh sieve and ground.After this, around 0.5 g from each sample was submitted to 10 ton cm -2 of pressure for 2 min to form pellets.For each soil sample, two pellets were prepared.Each pellet was placed in a sample holder with a quartz window and was measured using the benchtop and portable LIFS system.The portable LIFS system is also a lab-made equipment developed by Embrapa Instrumentantion at the Lasers and Optics Laboratory for this application (Figure 1).It comprises a diode laser (Coherent -CUBE) emitting at 405 nm (50 mW), an optical shutter, a bifurcated optical fiber bundle with seven optical fibers in a stainless steel ferrule: six illumination fibers around one read fiber (Ocean Optics), a high sensitivity mini-spectrometer (USB4000 -Ocean Optics), an adjustable optical filter, and a notebook.The resolution of the system was around 10 nm for all acquisition ranges (475-800 nm).In addition, a software was developed to control the laser, the shutter, and spectrometer parameters such as integration time and number of averages for each measurement.
The benchtop LIFS equipment is a lab-made system similar to that described by Milori et al. 36 In this paper, LIFS operation parameters were: (i) lock-in amplifier 100 mW, (ii) photomultiplier tube type of detector operating at 850 V, (iii) the laser used is an argon laser manufactured by Coherent; model Innova 90, tuned for emission at 458 nm blue region to excite soil fluorescence and (iv) 400 mW laser power.The spectral resolution was adjusted to 4 nm for all acquisition ranges (470-650 nm). 36or both LIFS equipment, the ratio between the area under fluorescence emission spectrum (FEA) (range 475 and 800 nm) and TOC concentration in the sample was considered as an indicator of humification of SOM (HD). 36he measurements were made in triplicate for each soil sample.The same soil pellets used in the benchtop system were measured to validate the portable system.

Comparison among methods
Fluorescence methods to assess humification degree of humic acid were compared: conventional fluorescence index 21,22,24 and laser induced fluorescence index. 36he humification degree of SOM assessed by portable LIFS system using whole soil was compared to humification degree of humic acid determined in solution by conventional methods of fluorescence spectroscopy.
Moreover, considering the benchtop LIFS system similar to that described by Milori et al., 36 the portable Performance Evaluation of a Portable Laser-Induced Fluorescence Spectroscopy System J. Braz.Chem.Soc.778 system was valued using basic statistical treatments.The calibration curve was fitted by the least squares method (LSM) using the HD obtained from the benchtop LIFS system as the independent variables and the HD obtained from the portable LIFS system as the dependent variables.The samples used to obtain the calibration model were the A1 to A30.For validation, the samples used were A31 to A36.

Results and Discussion
Organic carbon contents of samples are shown in Table 1.As can be seen, the used samples have low quantities of carbon, and the carbon content decreases with soil samples depth.Figure 2 presents typical spectra of fluorescence emission for whole soils and HA in solution.
For HA, it was possible to observe changes in spectral profiles and emission intensities.In the presented example, variations in intensity reached a factor of 5 (Figure 2a, 2b  and 2c).From the data of carbon obtained by TOC (Table 1) and the fluorescence spectrum area, the HD was determined for the diferent methods mentioned above.A comparison among methods to assess humification degree of HA is shown in Figure 3.The correlation among indexes (A 465 , A 4 /A 1 and I 468 /I 374 , Figure 3) can be considered very strong (R ≥ 0.95, with *p < 0.0001).
It is important to highlight that the three methods that use fluorescence are basically different from each other.While the method of Zsolnay et al. 22 uses ultraviolet excitation, where several groups of HA are excited, the method proposed by Milori et al. 24 excludes a great portion of the groups, being, therefore, more selective.Although the emission spectrum with excitation at 240 nm is wide (Figure 2a), it presents changes in the spectral profile, while the emission spectrum with excitation at 465 nm is narrower (Figure 2b), possessing the same spectral profile.When the synchronous-scan is carried out, Kalbtiz et al. 21ethod is also more selective than Zsolnay et al. 22 method, because it specifically shows a resonance whose difference between the excitation and emission is 55 nm (Figure 2c).
For analysis using whole soil, it was observed mainly changes in intensity of fluorescence emission.When excited at 458 nm, LIFS emission spectrum of whole soil samples shows a single broad band from 470 to 650 nm (Figure 2e), quite similar to those reported for HAs fluorescence spectra. 24hese spectra also show maximum fluorescence intensity around 520 nm, in all depths, increasing the fluorescence intensity in deeper depths.Other samples presented very similar spectra.Similarly, when excited at 405 nm (portable system), LIFS emission spectrum shows a single broad band from 475 to 800 nm (Figure 2d).It is interesting to emphasize that, for the presented spectra, variations in fluorescence intensity may reach a factor of 2. It means that the carbon structures are different.The quantum efficiency of fluorescence emission for samples A31 to A36 is bigger than for A7 to A12.In other words, there is a lot of carbon in samples A7 to A12, which are inside of unsaturated and rigid structures, like aromatic rings.
Figure 4 shows the correlation between the HD of humic acid determined by conventional fluorescence and the HD assessed by the portable LIFS system using pellets of whole soil samples.The Pearson's coefficients between HD determined using the portable system and the conventional fluorescence spectroscopy methods were high (R = 0.85 for A 465 ,  4a, 4b and 4c) and dispersion was low *p < 0.0001.Another interesting observation is the increase of dispersion in the HD of OM in deeper horizons (these are the horizons that have greater depth of soil samples and showed the highest humification degree) (Figure 4a, 4b and 4c).It is important to emphasize that the measurements using conventional fluorescence methods analyze only the humic acid and the experiments using LIFS analyze the total OM.Thus, the insoluble fraction of humic substances (humin) and fulvic acid are also analyzed by LIFS.In light of this, the humin fraction can comprise over 80% of humic material of SOM.
Intensity / a.u.Intensity / a.u.The correlations among the data are high and in both cases, the results indicate the feasibility of using the portable LIFS system as a tool for investigating the humification of organic matter in whole soil samples, without the need of chemical extractions.
Notwithstanding, the portable LIFS system is easier to use because the measurement is conducted using optical fiber, thus avoiding the misalignment that may occur with the benchtop LIFS system.The portable equipment is an analytical method able to provide a high analytical frequency, allowing direct analysis of soil samples meeting the demand for information on HD, which substantially reduces the scanning time. 43The correlation between the portable LIFS and the benchtop LIFS in estimating HD is shown in Figure 5a.The Pearson's coefficient (R = 0.953 with *p < 0.0001) indicates a very strong correlation.
The calibration model was fitted from the LSM using the HD determined by the benchtop LIFS system as independent variables and the portable LIFS system as dependent variables.The equation obtained from this model curve was H portableLIFS = [(66 ± 11) + (0.32 ± 0.02) (H benchtopLIFS )] × 10 3 (Figure 5a).The treatment samples used for the model validation were A31 to A36. Figure 5b shows the validation graphic, where the Pearson's correlation coefficient  (Milori et al. 24 ) and A 4 /A 1 (Zsolnay et al. 22 ) and (b) A 465 (Milori et al. 24 ) and I 468 /I 374 (Kalbitz et al. 21 ).was very strong (R = 0.997) with low dispersion factor (*p < 0.0001), indicating a low value of error in estimating the HD of SOM.Table 2 shows the HD values used in the calibration model and the values estimated by the portable LIFS system.As can be seen in Figure 5b and Table 2, the results from the portable LIFS system are very similar to the ones obtained with the benchtop LIFS system, thus demonstrating the great application potential of portable systems for the assessment of the HD of the SOM of whole soil samples.These results agree with those obtained by Segnini et al. 43 who used a portable LIFS system for assessing SOM stability in whole soil samples from permanent and seasonally flooded wetlands in the Peruvian Andes.Results from the portable LIFS showed significant correlation when compared with the benchtop LIFS system (R = 0.94), showing the feasibility of using a portable LIFS system and representing an improvement in speed and convenience for SOM stability evaluation across the landscape.
Coherence and concordance between the results obtained from the proposed calibration model and the HD estimated by the portable LIFS were observed; the relative error between them was lower than 11%.
The use of LIFS is based on the measurement of fluorescence of the whole soil through soil sample excitation with a laser in the violet or blue region (405-458 nm).This procedure privileged the fluorescence excitation of complex structures of OM associated with more recalcitrant part of HS (HA), such as aromatic and heterocyclic structures.Aliphatic structures should not show fluorescence signal in the violet or blue region as well as should not excite less humified structures of HS (fulvic acid).
It is important to report LIFS that generate semiquantitative information about the molecular complexity of SOM and thus provides a relative indication of its humification process.The equipment parameters are adjusted based on the strength of the signal produced by the whole soil samples and thus quantitative assessment are not feasible.Regarding the benchtop and portable LIFS systems used in this work, there were intrinsic differences of the experimental apparatus to generate specific signal intensity.Since spectroscopic techniques give no absolute humification degree, only relative comparisons are possible.The laser intensities used had to be different, for the obvious reasons; therefore, the electrical signals generated by those lasers are expected to be distinct. 43However, high correlation between systems that use whole soils and solutions is very interesting from an analytical point of view.This study can be a very interesting contribution for analytical chemistry, since this portable LIFS system is well suited to be applied in situ, enabling real time evaluation of soil quality through assessing of humification degree.

Figure 1 .
Figure 1.Experimental scheme of the portable LIFS system.

Figure 2 .
Figure 2. Fluorescence emission spectra of HA obtained with excitation in 240 nm (a), excitation in 465 nm (b), synchronous scan using a Δλ of 55 nm (c) and fluorescence emission spectra of whole soil samples obtained by the portable LIFS (d) and by the benchtop LIFS systems (e) for the sample less humified (A31 sample: solid lines) and more humified (A12 sample: dashed lines).

Figure 5 .
Figure 5. (a) Calibration model for portable LIFS using the benchtop LIFS systems already described in the literature by Milori et al. 36 for samples: A1 to A30.(b) Validation of the model using samples A31 to A36.

Table 1 .
Carbon concentration of whole soil samples obtained by the elemental analysis (TOC) R = 0.83 for A 4 /A 1 , and R = 0.80 for I 468 /I 374 , Figure

Table 2 .
HD estimated by the calibration model and estimated from the portable LIFS system using the simple linear regression model for the sample of A31 to A36