Meat-Processing Wastewater Treatment Using an Anaerobic Membrane Bioreactor (AnMBR)

Line

Comment

Ensure all acronyms are defined on first use, and first use only. Some acronyms currently not defined.

Figures should appear immediately following their first mention in the text.

Ensure all reference items appear in the text, and all citations in the text appear in the reference list.

Please streamline the content for clarity. For example, COD removal is discussed a long time after biogas yields per unit of VS removed. Here we are left asking how much COD was not removed? – and this distracts from your messaging. Re-ordering content would help improve readability.

There are numerous typographical errors which require attention.

Abstract

The abstract should be a total of about 200 words maximum

keywords

Keywords currently mirror title. Please use synonyms or alternatives to title words to expand searchability. N.B. both title words and keywords are used by search engines to find results.

17

g/(kg*d) is an unusual way represent OLR. I assume the authors are assuming a density of 1 and are substituting kg for litres. Note that OLR is a mass per volume per time measurement.

18

Maximum methane yield was “almost” 350 Nm3/t COD

30

Revise keywords to be different to title words to boost searchability.

58

Red meat processers have typically overcome the issue of high volumetric load by constructing large covered anaerobic lagoons – See America and Australia. These lagoons are typically in the order of 20+ megalitres and, due to their size, enable solids to sediment out and to some degree this decouples the solids retention time from the hydraulic retention time. Such systems also rely on passive heating and, consequently, I suspect it is these issues which make covered anaerobic lagoons unsuitable for the Austrian context, and not simply a high volumetric load.

59-66

Paragraph is incoherent. Topic sentence is about reactor designs to overcome high volumetric loads, but only introduces UASB and quickly changes topic to freshwater recovery from wastewater. The paragraph could allude to UASB, EGSB, AFR, AnMBR, baffled reactors, fluidised bed reactors, etc.

86

The justification for the research “While there 84 have been improvements in wastewater treatment within this sector, the industry's high 85 energy and water demands necessitate further innovations [11].” comes from 1995 – some 30 years ago. Substantial work has been conducted in this space since 1995, albeit the sentiment of the statement remains true. Recommend updating to a more recent reference.

89-92

Please further elaborate on the novelty of the study.

Further to this, there is a distinct lack of citations which include AnMBR use in treating slaughterhouse waste, of which there are several published investigations.

A select few include:

Jensen et al. (2015) explored AnMBRs for treating red meat processing wastewater https://doi.org/10.1016/j.bej.2015.02.009

Viet, Vu and Duong (2023) explored the effect of HRT on AnMBR performance treating slaughterhouse waste. https://doi.org/10.1016/j.envres.2023.116522

Gautam et al. (2023) evaluated long-term operational and treatment performance of an AnMBR treating slaughterhouse wastewater. https://doi.org/10.1016/j.cej.2023.142145

111

Include CSTR volumetric capacity.

155

Inhibition testing…

158

Acronyms already defined.

158-174

A new paragraph is not needed for each analysis.

288-310

Combine into a single paragraph. Provide concluding statement about the impact of salinity in the CSTR and inhibition. What insights do you draw by determining the anion composition of the ash?

313

Provide more insightful topic sentences for your paragraphs.

314-5, 317-8

Move to methods.

323-338

This is difficult to follow:

1.       BMPs were conducted in which two inocula were combined at a ratio of 1:1 to reach a TS of 3.5%. Was 3.5% TS a target? Or was this just a coincidental result of 1:1 addition that it exactly reached 3.5%? Is the 1:1 ratio on a wet mass basis? TS basis? VS basis? This content should be explained in the methods, not in the results.

2.       You determined a methane potential of 635 Nm3/t VS (see unit error kNm3, line 327), and indicated that the yield was greater than estimated. You’ve then indicated that the inoculum contained significant quantities of fibrous material. Was the inoculum screened/sieved prior to use? Were positive and negative control run alongside the trials to a) determine test validity, and b) subtract inoculum gas production from substrate gas production?

3.       What inoculum to substrate ratio (ISR) was used? Low ISR can result in outcomes that are difficult to explain, especially extraordinarily low BMP yields. 2:1 on a VS basis is stock standard, with 3:1 and higher more common in high-fat substrates or potentially inhibitory substrates.

4.       What was the standard deviation of the BMP values?

5.       Which methodology was followed for the inhibition assay? Which inhibitors did you assess?

6.       The conclusion of the sludge not being adapted to the wastewater is not well supported. More likely that the inoculum to substrate ratio was insufficient. How long did the BMP Take? The presence of a cellulose control would have helped to clarify the situation.

350-351

Figure 2 indicates several inconsistencies in HRT/feed rate and is attributed to “maintenance activities, feed exchanges and membrane cleaning procedures”. The paper glosses over the impact of membrane fouling. As indicated by Jensen et al. (2015), “Optimisation and control of membrane fouling will be a core area for ongoing research and development.” Fundamentally, operation of wastewater systems are outside core business of red meat processors and they are generally disinterested in high-maintenance infrastructure – even that which carries an economic case to employ a full time technician. Could the authors please elaborate on membrane fouling in their experiments?

Figure 3

Regarding the y axes, the left axis is in m3/d, while the right axis is in g/L/d. You’re dealing with an OLR of 0-5.5g COD/L/d. Volume of CH4 is largely irrelevant if reactor volume is unreported. Please both report reactor volume (in the methods) and change the left axis to be methane production in NmL/L/day

385-393

It is unclear why the authors have decided to report specific methane yield on the portion of organics removed, rather than loaded which is far more common. Indeed, there are works which report the yield per unit of organics removed, however, this artificially inflates the specific methane yield by ignoring the undegraded organics. Essentially, this ignores the degradable fraction of COD or VS which should be determined as part of BMP testing and cannot be accurately estimated under continuous conditions. For instance, in figure 5 phases 2 & 3, m3 CH4/t COD is reported to have exceeded the theoretical maximum of 350 m3 ch4/t COD at times substantially.

For transparency, please convert specific methane yield to volume of ch4/ mass of organics loaded.

398-408

Biogas and methane theoretical yields were estimated by determining the relative proportion of protein, carbohydrate and fat, and multiplying these proportions by their theoretical maxima as per the Buswell equations or similar. This assumes a degradable fraction of 100%. The authors have then compared this with their results in which the workup of their data also assume 100% degradation (through subtraction of undegraded organics), but they do not actually achieve 100% degradation. This essentially prioritises the rapidly degradable organics and may artificially influence the specific methane yield. Please amend to either state the specific methane yield per unit of organics loaded or include the degradable fraction for each phase.

507-548

It is unclear how the microbial community study contributes to the research aim. Please clarify.

550-576

Conclusion contains a rehash of results and discussion that in the opinion of this reviewer is too lengthy and could be more succinctly presented in a response to the research question/aim.

The claim that substrate composition influenced the microbial community was not explored in this study. The lack of focus on the microbial community study emphasises that lack of purpose for the investigation in this work.

Supplementary

Please provide inhibition study as supplementary materials.

Line

Comment

Include % VS/TS or %VS/FM for each inoculum used.

188-190

CSTR had a working volume of 11 litres and a head space of 2L, but a total sludge volume of 18.6L? how is this possible?

185

Inoculum to substrate ratio (ISR), or substrate to inoculum ratio (SIR)? 0.18 g VS from inoculum per 1 gram of VS of substrate is extremely low. This should be in the order of 2+. However, substrate to inoculum ratio is the inverse of this and consequently should be <=0.5. Please check throughout.

346

Word missing “Pose a by”

369

Check Table 3. By my estimation, assuming a COD/VS ratio of 1.07 for carbohydrates, 1.42 for proteins, and 2.88 for fats, and using your substrate VS composition of 13.6% carbohydrate, 60.8% protein and 25.6% fat, the substrate has a COD load of 1.74616 g COD/g VS. As the theoretical maximum methane yield per gram of COD is 350 mL, your theoretical maximum methane yield is 611 ml/g VS. Using VDI 4630 (2016), the fractions of methane from carbohydrates, proteins, and fats are 50%, 50% and 72% respectively, giving a total theoretical biogas yield of 1064.6 mL/g VS. Your methane yield of 272 m3/t COD indicates a degradable fraction of 78%.

373-374

The variation between your inocula is still interesting, as figure A3 indicates that degradation was rapid and generally unimpeded in both instances. Was the exact same batch of substrate used for each test? Fat is a very difficult material to work with as it clumps in cold storage and becomes difficult to reproducibly dispense, leading to large variation in results, and in general it floats on the surface of a reactor and can escape digestion by adhering to the side of the vessel. Because of this it is possible that varying amounts of residue were subject to digestion.

382-384

I am more convinced that the ‘standard inoculum’ contained additional organics that were released during digestion. I understand de-gassing was conducted to consume residual organics which explains the relatively consistent BMP testing. However, BMP testing in which the AnMBR inoculum was used are all monophasic and show rapid degradation of a relatively simple substrate. However, the BMP testing using ‘standard inoculum’ which we know contained lignocellulosic material, degraded with bi- and even almost tri-phasic behaviour which suggests more complex degradation is occuring. I suspect that some degree of co-digestion or additional breakage of ligno-cellulosic bonds is occurring under the more acidic conditions induced by digestion and liberating additional organics for biogas production.

401-402

Not a paragraph. Combine with paragraph at 406. Move figure below text.

485

‘Extend’ should be ‘extent’.

489-491

“The result of 519 Nm3/t(VS removed) is again close to the yield of the adapted AnMBR inoculum batch test of 498 Nm3/t(VS) and much higher compared to the yield of loaded COD reaching 398 Nm3/t(VS).”

It is unclear what you mean by the yield of loaded COD reaching 398 m3/tVS. Is it COD or VS? COD cannot reach 398 due to a maximum of 350.

495-498

“For the comparability to other systems, the yield per loaded organics remains the relevant parameter, even though the [yield] of the BMP test results are better comparable to the yield per removed contaminants.”

Yes, but this should always be true. As the BMP is going to completion, the amount of undegraded organics is minimised and VS removed comes closer and closer to VS loaded. If you focus only on the organics removed, you should always achieve around theoretical max which will be most closely achieved by your BMP test. There is a good reason that the literature focuses on yield per unit of organics loaded and not just those removed, though the degradable fraction is still an important substrate parameter.