Experimental and Numerical Study of Biochar Fixed Bed Column for the Adsorption of Arsenic from Aqueous Solutions

Round 1

Reviewer 1 Report

The authors describe a numerical model that they have applied to a laboratory test to study the retention capacity of As by an adsorbent material.

The experiment has been carried out with different conditions that have been explained in the text

However, they should explain why they have used Hazen's expression to estimate permeability, when they could have obtained it using a darcy test. The poor fit obtained between the numerical model (Fig. 3 and 4) and the experimental results may be due to the fact that the advective flow is greater in the model than the diffusive one. What is the value of Diffusivity?

They should correctly estimate the value of permeability and validate it with a previous model.

Author Response

Response to Reviewer 1 Comments

The authors describe a numerical model that they have applied to a laboratory test to study the retention capacity of As by an adsorbent material.

The experiment has been carried out with different conditions that have been explained in the text

Point 1: However, they should explain why they have used Hazen's expression to estimate permeability, when they could have obtained it using a darcy test. The poor fit obtained between the numerical model (Fig. 3 and 4) and the experimental results may be due to the fact that the advective flow is greater in the model than the diffusive one. What is the value of Diffusivity?

 Response 1: Thank you for your suggestions. The use of the Hazen’s equation was due to the fact that the permeability was not used in the model, the reported value was just used to give a complete view of the materials used in the tests. Concerning the “poor fit obtained between the numerical model and the experimental results” we disagree slightly with the reviewer because it is possible to underline that the breakthrough curves detach well the origin of the axes, indicating clearly that the value of the center of gravity of the contamination plume is optimally localized by means of a correct definition of the seepage velocity. The low fitting is localized in the terminal part of the curves, this is due to a lack in the information collected during the tests.

The calibrated value of the hydrodynamic dispersion (D=5*10^-5 m² min^-1) reported in the Table 3 is the sum of two components: mechanical dispersion and diffusivity. The magnitude of arsenic diffusivity in water was found to be negligible (≈10^-8 m² min^{-1 **}), so the main contribution came from the mechanical dispersion (≈ 10^-5 m² min^-1).

 ^** Tanaka, M., Takahashi, Y., Yamaguchi, N., Kim, K. W., Zheng, G., & Sakamitsu, M. (2013). The difference of diffusion coefficients in water for arsenic compounds at various pH and its dominant factors implied by molecular simulations. Geochimica et Cosmochimica Acta, 105, 360-371.

Point 2: They should correctly estimate the value of permeability and validate it with a previous model.

Response 2: Thank you for the point! As reported in the answer to point 1, the permeability of both materials used during the test (biochar and sand) are not used in the numerical code; therefore the values were not considered as an important aspect of the work. As already said we reported, just for information purposes, in Table 2 the permeability values of the sand.

Author Response File: Author Response.docx

Reviewer 2 Report

Following an overall inquiry into the reviewed article, I consider it to be a very interesting investigation of experimental and numerical study of biochar fixed bed column for the adsorption of arsenic from aqueous solutions.

The manuscript contains interesting and valuable data, which have been mostly correctly evaluated and interpreted. Organization and clarity of the manuscript is also generally good.

The paper resolves an elaborate multidisciplinary topic and meets formal layout standards and default criteria, imposed on such articles.

Thereby, I recommend its issuance.

I would like to ask the authors to marginally mention the following article in the introduction: V. Frišták et al.: Sorption separation of Eu and As from single‐component systems by Fe‐modified biochar: kinetic and equilibrium study. J. Iran. Soc. 14, 521 (2017). E. Viglašová et al.: Production, characterization and adsorption studies of bamboo-based biochar/montmorillonite composite for nitrate removal. Waste Management. 79, 385 (2018). M. Daňo et al.: Surface Complexation Models of Pertechnetate on Biochar/Montmorillonite Composite—Batch and Dynamic Sorption Study. Materials. 13, 3108 (2020).

Author Response

Response to Reviewer 2 Comments

Following an overall inquiry into the reviewed article, I consider it to be a very interesting investigation of experimental and numerical study of biochar fixed bed column for the adsorption of arsenic from aqueous solutions.

The manuscript contains interesting and valuable data, which have been mostly correctly evaluated and interpreted. Organization and clarity of the manuscript is also generally good.

The paper resolves an elaborate multidisciplinary topic and meets formal layout standards and default criteria, imposed on such articles.

Point 1: Thereby, I recommend its issuance. I would like to ask the authors to marginally mention the following article in the introduction: V. Frišták et al.: Sorption separation of Eu and As from single‐component systems by Fe‐modified biochar: kinetic and equilibrium study. J. Iran. Soc. 14, 521 (2017). E. Viglašová et al.: Production, characterization and adsorption studies of bamboo-based biochar/montmorillonite composite for nitrate removal. Waste Management. 79, 385 (2018). M. Daňo et al.: Surface Complexation Models of Pertechnetate on Biochar/Montmorillonite Composite—Batch and Dynamic Sorption Study. Materials. 13, 3108 (2020).

Response 1: Thank you for your kind suggestion. References suggested by the reviewer have been added to the manuscript.

Line 79. “… so biochar may represent a sustainable material [53–55].”

53. Viglašová, E.; Galamboš, M.; Danková, Z.; Krivosudský, L.; Lengauer, C.L.; Hood-Nowotny, R.; Soja, G.; Rompel, A.; Matík, M.; Briančin, J. Production, characterization and adsorption studies of bamboo-based biochar/montmorillonite composite for nitrate removal. Waste Manag. 2018, 79, 385–394, doi:10.1016/j.wasman.2018.08.005.
54. Frišták, V.; Micháleková-Richveisová, B.; Víglašová, E.; Ďuriška, L.; Galamboš, M.; Moreno-Jimenéz, E.; Pipíška, M.; Soja, G. Sorption separation of Eu and As from single-component systems by Fe-modified biochar: kinetic and equilibrium study. J. Iran. Chem. Soc. 2017, 14, 521–530, doi:10.1007/s13738-016-1000-1.
55. Daňo, M.; Viglašová, E.; Galamboš, M.; Štamberg, K.; Kujan, J. Surface Complexation Models of Pertechnetate on Biochar/Montmorillonite Composite—Batch and Dynamic Sorption Study. Materials (Basel). 2020, 13, 3108, doi:10.3390/ma13143108.

Author Response File: Author Response.docx

Reviewer 3 Report

Arsenic adsorption onto different kinds of adsorbents in a century old research (Huang, C. P., Fu, P. L. K., Treatment of arsenic (V)-containing water by the activated carbon process. J. Water Pollution Control Federation, vol. 56, no. 3, part 1, pp 233-242, 1984; Huang, C.P., Vane, L. M., Enhanced As5+ removal by a Fe2+-treated activated carbon, J. Water Polution control Federation. Vol. 61, no. 9 p 1596-1603, 1989). Arsenic can present in most natural waters in As(III) and As(V) forms. AS(V) was used in the study. As(V) is a weak Bronsted acid with three acidity constants, pKa1 = 2.19, pKa2 = 6.94 and pKa3 = 11.5. Furthermore both quartz sand and biochar will be charged also. Therefore, pH is an important parameter controlling arsenic adsorption. Authors have ignore the effect of pH on As(V) adsorption.

Table 1 lists the As adsorption capacity of different absorbents. However the data are not useful, because the conditions at which the adsorption capacity reported were not given. Note that the adsorption capacity can be affected by the pH, ionic strength and temperature to name a few.

Adsorption columns were packed with biochar mixed with “inert quartz sand”.  In fact, quartz sand is not as inert as authors have indicated as shown in the results presented  (Fig. 2).

What kind of material was used to prepare the column? Plexi glass?

Important physical chemical properties such as surface charge and ash content were not characterized. How was the electrical conductivity of biochar determined? What is it has anything to do with As adsorption?

Design of column reactor was unnecessarily excessive, that is, there is not to take samples throughout the height of the column; only port D is needed. Results showed in Fig. 2 indicated that test 1 breaks through at 8 hours and test 2 breaks at 10 hours. This indicates that column 2 has greater As adsorption capacity than column 1. However it is noted that column 1 has more biochar (7:100) than column 2 ( 3:100). Results indicated that biochar does not contribute significantly As adsorption capacity!!!!  That is, As adsorption in both columns actually is contributed by quartz sand. Question, is quartz sand really “inert” as authors have stated in the Introduction section?

The kinetics of ion adsorption can be mass-transfer or surface reaction controlled. Equation predicts mass-transfer controlled adsorption. There is an “adsorption zone” developed in the column. However, the adsorption profile appears to belong to “surface reaction” controlled. In this case, the Adam-Borhart equation should be applied instead of Equation 3.  Table 3 lists the value of model-fitting parameters. What is “P”?  Both Kd and KL values are listed in Table 3. But Kd is applicable at low As concentration. Are Kd and KL used at the same time to model the As column adsorption?

Author Response

Response to Reviewer 3 Comments

Point 1: Arsenic adsorption onto different kinds of adsorbents in a century old research (Huang, C. P., Fu, P. L. K., Treatment of arsenic (V)-containing water by the activated carbon process. J. Water Pollution Control Federation, vol. 56, no. 3, part 1, pp 233-242, 1984; Huang, C.P., Vane, L. M., Enhanced As5+ removal by a Fe2+-treated activated carbon, J. Water Polution control Federation. Vol. 61, no. 9 p 1596-1603, 1989). Arsenic can present in most natural waters in As(III) and As(V) forms. AS(V) was used in the study. As(V) is a weak Bronsted acid with three acidity constants, pKa1 = 2.19, pKa2 = 6.94 and pKa3 = 11.5. Furthermore both quartz sand and biochar will be charged also. Therefore, pH is an important parameter controlling arsenic adsorption. Authors have ignore the effect of pH on As(V) adsorption.

 Response 1: We would like to thank the Reviewer for the note. The reagent used for the preparation of the aqueous solutions is an As (V) salt; however, the analytical method provides for the determination of total arsenic concentrations in the aqueous phases, therefore it was not possible to determine the arsenic speciations present and the related oxide reduction phenomena. We supplemented the text, specifying that the analysis concerned the total arsenic concentration, as reported below.

Even if it was well known the importance of pH on the behavior of the investigated element we did not determine the pH in the reported samples during testing; we only determined it at the beginning because our main task was to verify the efficiency and the potentiality of the adsorbent material on As. Therefore the topics on which the work has been focused were mainly the possibility to use BIOCHAR as adsorbent and the amount necessary to reach defined efficiencies, it is already planned to carry out specified lab tests to check the influence of other physical and chemical parameters on the overall results.

These preliminary tests have allowed to estimate the influence of dosage to be used so to have basic information in the next test, in which we will investigate further the optimal operating conditions.

In any case, since the tests were carried out in a closed system, there has been no exchange with the CO₂ that could partially change the initial measured pH.

We are fully agreed with you and we will consider this great advice in future experimental work.

Line 141-146: “Total Arsenic concentration in the aqueous phases was determined using mass spectrometry with inductive plasma source (ICP-MS; Perkin-Elmer®, Model NexION 300x), whose detection limit was 1 µg/L.

The calibration curve was determined using Standard Methods-3125 [69] and total arsenic solutions at four concentration (0, 10, 50 and 100 µg/L As).”

Point 2: Table 1 lists the As adsorption capacity of different absorbents. However the data are not useful, because the conditions at which the adsorption capacity reported were not given. Note that the adsorption capacity can be affected by the pH, ionic strength and temperature to name a few.

Response 2: We would like to thank the Reviewer for the remark. As requested by the reviewer, Table 2 has been modified, reporting all the available data of the investigated materials, in particular the pH of the contaminated solution, the average size and the specific surface of the investigated materials.

Point 3: Adsorption columns were packed with biochar mixed with “inert quartz sand”.  In fact, quartz sand is not as inert as authors have indicated as shown in the results presented  (Fig. 2).

Response 3: We thanks the reviewer for the comment. In Figure 2, breakthrough curves are reported in both tests, using different biochar volume ratios within the experimental setup. An inert quartz sand was used as filling material of the columns to avoid contributions related to the possible adsorption of arsenic on the sand, to avoid packing phenomena and to simulate a possible reactive permeable barrier at laboratory scale. In fact, at field scale reactive permeable barrier are filled with adsorbent materials and with inert materials, as quartz sand.

Point 4: What kind of material was used to prepare the column? Plexi glass?

Response 4: Thanks for the comment. To avoid interaction phenomena between the reactive material and the column, an experimental apparatus in Pyrex glass was deliberately used.

Point 5: Important physical chemical properties such as surface charge and ash content were not characterized. How was the electrical conductivity of biochar determined? What is it has anything to do with As adsorption?

Response 5: Thanks for the comment. The ash content was reported in Table 2, as required. As far as the specific surface is the subject of our recent studies, we hope to present it as soon as possible. Electric conductibility of biochar was determined using the UNI EN 13040: 2008 + UNI EN 13038: 2012, as described in Table 2. Unfortunately the electrical conductivity was presented for information in the technical data sheet and was not determined during the test, as we recall that the main objective was to verify the optimal dosage and the numerical model implemented to validate the experimental data collected.

We are fully agreed with you and we will consider this great advice in our recent and future experimental work.

 Point 6: Design of column reactor was unnecessarily excessive, that is, there is not to take samples throughout the height of the column; only port D is needed. Results showed in Fig. 2 indicated that test 1 breaks through at 8 hours and test 2 breaks at 10 hours. This indicates that column 2 has greater As adsorption capacity than column 1. However it is noted that column 1 has more biochar (7:100) than column 2 ( 3:100). Results indicated that biochar does not contribute significantly As adsorption capacity!!!!  That is, As adsorption in both columns actually is contributed by quartz sand. Question, is quartz sand really “inert” as authors have stated in the Introduction section?

Response 6: Thanks for the comment. Anyway, we disagree with the first part of the comment reported from the reviewer. The use of a numerical model need, especially in the calibration and validation phase, of the most possible information. The intermediate sampling ports are extremely useful for the check also of the dispersive parameter used. So, the several ports used to sample have helped us to verify the behaviour of the whole all along the column. Through these first tests, in which the volume ratio was varied in order to verify the relative adsorption capacity, we found that the second column reaches a better performance. The reasons of this unexpected behaviour are surely due to the particular characteristics of the BIOCHAR, it is an extremely light material with particle dimensions really small. These characteristics brings to an inhomogeneous dispersion of the solution and to zones (in this case into the column) in which the deep penetration and therefore the contact between the adsorbent material and the solution could be not enough or totally absent. In this situation  not all the amount of adsorbent material participates to the removal process and the efficiency decrease, therefore, as for other material which are tested for a future use in groundwater treatment (especially if in situ), the evaluation of the correct ratio between the active material and the inert one must be studied and tested.

Point 7: The kinetics of ion adsorption can be mass-transfer or surface reaction controlled. Equation predicts mass-transfer controlled adsorption. There is an “adsorption zone” developed in the column. However, the adsorption profile appears to belong to “surface reaction” controlled. In this case, the Adam-Borhart equation should be applied instead of Equation 3.  Table 3 lists the value of model-fitting parameters. What is “P”?  Both Kd and KL values are listed in Table 3. But Kd is applicable at low As concentration. Are Kd and KL used at the same time to model the As column adsorption?

Response 7: We would like to thank the Reviewer for the interesting remark. A numerical model was deliberately used, rather than an analytical one like the one mentioned by Bohart – Adams, mainly due because of its greater flexibility in simulating the experimental data collected, in order put the basis for  future studies on the use of this material at the real scale. P is the porosity (Line 162).

Author Response File: Author Response.docx

Round 2

Reviewer 1 Report

The revision is attached to the file 

Comments for author File: Comments.pdf

Author Response

Point 1: The revision is attached to the file.

Response 1: Thank you for your comment. We agree to the Reviewer regard the dependence of the mechanical dispersion (D) on the porosity and on the Darcy velocity. The calibrated value of D, reported in the Table 3, includes the effects of the parameters mentioned above. As explained in the paper, the used numerical model is one dimensional, thus D is expressed as:

                                                                                                                             (1)

where u is the seepage velocity and a_L the longitudinal dispersivity.

U was calculated using the following equation:

                                                                                                                             (2)

where Q is the solution flow rate (equal to 5 mL/min), A the column section and p the porosity. The estimated value of a_L is 1.3*10^-2 for the Test 1 and 1.5*10^-2 for the Test 2. The obtained values are in accordance with the values estimated using the relation proposed by Pickens and Grisak, 1981. In lines 161-166 we have described the calculation of the seepage velocity and in Table 3 we have added the values of u for the two tests.

Pickens J. F., Grisak G. E. Scale-Dependent Dispersion in a Stratified Granular Aquifer. Water Resources Research, vol. 17, 4, 1191-1211, 1981.

Line 161–175:

The motion in the column was assured from a pump connected to the exit of the column, in this way the seepage velocity was calculated from the equation (2):

where Q is the solution flow rate (equal to 5 mL/min), A the column section and p the porosity.

The value of the hydraulic dispersion coefficient (that coincides with mechanical dispersion), D, was calibrated and the dispersivity value was determined for the two tests by means of:

The estimated value of a_L is 1.3*10^-2 for the Test 1 and 1.5*10^-2 for the Test 2. The obtained values are in accordance with the values estimated using the relation proposed by Pickens and Grisak [70].

Author Response File: Author Response.docx

Reviewer 3 Report

Authors have not totally addressed concerns raised initially. Once authors have taken care of the following, the manuscript can be accepted for publication at Water. 

1. Line 96, what form of arsenic chemical was used in the study? Is it As(V) or As(III).
2. Add discussion on the speciation of As(V) or As(III) as a function of pH. That is, what is the major As(V) or As(III) species involved in the reaction?   
3. Line 127. delete "inert".  Quartz sand is not "inert" toward oxyanions! 
4. Line 121. Pyrex glass column was used. What is the amount of As(V) or As(III) retained by the glass wall? Pyrex glass surface can adsorb oxyanions!
5. Add table to give the fitting variables for Figure 3 and figure 4. 

Author Response

Point 1: Line 96, what form of arsenic chemical was used in the study? Is it As(V) or As(III).

Response 1: We thanks the reviewer for the comment. The arsenic solution used for the tests was obtained from a stock solution containing 3.125 g/L of As(V). It was prepared by dissolving Sodium arsenate heptahydrate (Na₂HAsO₄.7H₂O)

Line 110–113.

The arsenic solution used for the tests was obtained from a stock solution containing 3.125 g/L of As(V). It was prepared by dissolving Sodium arsenate heptahydrate (Na₂HAsO₄^.7H₂O) [62], which has a solubility of 39 g/100 mL into Milli-Q water at temperature (T) of 21 ± 0.1 °C.

Point 2: Add discussion on the speciation of As(V) or As(III) as a function of pH. That is, what is the major As(V) or As(III) species involved in the reaction?

 Response 2: We would like to thank the Reviewer for the note; we agree with this integration in the text.

Obviously, Redox potential (ORP) and pH are the most important factors controlling arsenic speciation. In order to control any oxidation-reduction phenomena, the pH of the solution was kept constant at 7.5; moreover, in all the samples collected during all the tests, the pH and the ORP were constant, respectively, 7.5 ± 0.2 and 800 ± 100 mV; therefore according with (Cullen and Reimer 1989), it is evident that arsenates (HAsO₄^2–) remained dominant. Therefore it is possible to show that As(V) is the major species involved in the reaction.

Cullen, W. R., & Reimer, K. J. (1989). Arsenic Speciation in the Environment. Chemical Reviews, 89(4), 713–764. https://doi.org/10.1021/cr00094a002

Line203–207.

The pH along with the redox potential (ORP, mV) are the most important factors for controlling the speciation of arsenic [76].

In fact, within the samples collected during all the tests, the pH and the ORP were constant, respectively, 7.5 ± 0.2 and 800 ± 100 mV; therefore, it is evident that arsenates (HAsO₄^2–) remained dominant [77].

Point 3: Line 127. delete "inert". Quartz sand is not "inert" toward oxyanions!

 Response 3: We thanks the reviewer for the comment. We corrected as requested by the reviewer.

 Point 4: Line 121. Pyrex glass column was used. What is the amount of As(V) or As(III) retained by the glass wall? Pyrex glass surface can adsorb oxyanions!

Response 4: We thanks the reviewer for the comment.

Considering the value of the initial concentration (1 mg/L) of the solution of As (V) and the mass balances calculated, the amount of arsenic retained by the column was considered negligible.

 Point 5: Add table to give the fitting variables for Figure 3 and figure 4.

Response 5: We thanks the reviewer for the comment. Table 3 has been inserted with the all parameters used for fitting the breakthrough curves of both column tests.

Line 250:

Table 3. Values of the parameters used for the simulation of the two column tests.

Please see the attachment.

Author Response File: Author Response.docx

Round 3

Reviewer 1 Report

Suggetsions: The authors could introduce the explanation given in the text, since it was not clear

Reviewer 3 Report

Authors have addressed most of the comments raised, except the pH and the reactor kinetics issues. Obviously, authors still do not understand the chemistry of arsenic in aqueous solution. Further, authors seem to lack understanding  the fundamental of reactor kinetics. Nonetheless, authors have shown some useful result on the removal of total arsenic by biochar.