*5.2. Impact of Ni-Based Ions and NP Addition*

The impact of Ni-based ions and NPs upon HY and HER is summarized in Table 2 and the statistical analysis results are shown in Figure 4.

**Table 2.** Comparison of BioH2 production with the addition of Ni-based nanoparticles.


In this table, MEG refers to mono ethylene glycol, CS: cornstalk.

**Figure 4.** Statistical analysis of HY and HER. (**A**) Particle size and nanoparticle concentration versus HY, (**B**) particle size and nanoparticles concentration versus HER.

Among the collected literature reports, the size and concentration of NPs together with their combined effect were not statistically significant to either HY or HER according to the calculated *p*-value. Regarding HY (Figure 4A), it was found that both too low and too high levels of NPs size and concentration were not favorable. Indeed, an optimal range existed if the NP size and concentration were manipulated within 86–120 nm and (81–120 mg L<sup>−</sup>1, respectively. Similarly, the HER also presented the same variation patterns as those of HY. An optimal range of particle size (86–120 nm) and Ni-ion/NPs concentration (81–120 mg L<sup>−</sup>1) existed for HER. Unlike Fe, Ni presented more consistent responding patterns between HY and HER in regards to the variation in the size and concentration of NPs. In addition, studies have indicated that Ni-based ions and NPs tend to selectively enhance some BioH2 generation pathways, such as enhancing the acetate pathway while suppressing or inhibiting butyrate and propionate pathways. However, discrepancies still exist due to different strains of microbes inoculated, cultivation medium, experimental uncertainties, etc. Although the size of NPs was significant to the HER, the combined effects (NP size and concentration) were found to be insignificant. Among the collected literature reports, the HER seemed to be directly related to the relatively larger size of the particles. This indicates that the manipulation of NPs ideally in size range of 81–100 nm is favorable for both HY and HER. This might contradict the first impression that the reduction of NPs size significantly enhances the quantum dot effect that subsequently boosts electron transport. However, the preparation and large-scale deployment of small-sized NPs that can stably exist in the cultivation medium has always been a substantial challenge, which will inevitably increase fixing and operating costs. Fortunately, the enhancement of BioH2 generation seems to be linked to an ideal range of NPs at the size of 81–100 nm; therefore, blindly pursuing small nanoparticles may be meaningless.

#### *5.3. Impact of Other Metal and Non-Metal Nanoparticle Addition*

The impact of other metal and non-metal NPs addition upon BioH2 generation is summarized in Table 3.

**Table 3.** Comparison of BioH2 production with the addition of other nanoparticles, where POME: palm oil mill effluent.


The addition of NPs was found to be effective at improving BioH2 generation due to the fact that NPs can facilitate electron transport in extracellular cultivation medium during fermentation [66,67]. With regard to the HY and HER, it was quite hard to find one individual NPs that positively enhanced both HY and HER simultaneously. This reflects the complex features of the BioH2 generation process, which generally involves many different steps of sub-metabolic pathways [43,68]. Among the investigated collected literature, CoO-NPs addition was among the most appreciable enhancement for HY and Ag-NPs addition was the most influential factor for HER enhancement. In addition, the impact of adding NPs prepared from hybrid approaches such as combining two different kinds of NPs, i.e., Cu and SiO2, was marginal. The correlation between BioH2 generation values (HER and HY) and the corresponding size of the NPs added to the fermentation broth was constructed and is plotted in Figure 5. The corresponding HY and HER varied from 0–30 (mmol g<sup>−</sup>1) and 0–80 (mmol L<sup>−</sup><sup>1</sup> h<sup>−</sup>1), respectively. Regarding to the enhancement of HY, some reported that smaller size (less than 42 nm) surely increased HY from 10 to 20–25 mmol g<sup>−</sup>1. On the other hand, for the enhancement of HER, some reported that a relatively bigger size of 40–50 nm seemed to significantly increase the H2 evolution rate. However, by considering the numbers of reports, the majority of works showed (i) the size of NPs seems to be more effective in enhancing HY than HER, and (ii) the rate of H2 evolution seems to be less responsive to the size of NPs, though some literature reported exceptionally higher values of HER after NPs (40–50 nm) addition.

**Figure 5.** Impact of size of NPs upon HER and HY.

#### *5.4. Impact of Ion Addition*

In this work, in order to assess the concentration impacts of different ions upon HY and HER BioH2 generation, ions including Mg2+, Cu2+, Na+, NH4 +, and K+ were selected and all data are summarized in Table S1.

It is worth noting that some metal ions inevitably introduced into the culture medium due to the use of NPs addition are not in the scope of discussion. It was quite challenging to find out the detailed concentration ranges in each study due to the factor that many reports did not specify the detailed cultivation steps. Although this could be difficult for estimating the level of those ions during the cultivation, the type of defined and undefined cultivation media used in the studies could be utilized to indirectly estimate the range of those different ions accordingly. The level of different ions upon HY and HER BioH2 generation are summarized in Tables S2 and 4, respectively, and the collected values from the literature were statistically analyzed through our previously established ANNs-RSM method.

By comparing the p-values, the impact of the variations in ion concentrations upon HY and HER of BioH2 generation could be identified accordingly [16,69]. It was found that the variations in the investigated ions only statistically influenced HER, but not HY. This suggests pivotal guidance for process intensification for BioH2 generation. The manipulations of ion concentrations in cultivation media can effectively improve or inhibit the rate but not the potential limit of BioH2 generation. In other words, the kinetics of BioH2 generation can be altered by varying some level of ionic concentration. The statistically significant impact of metal ion addition on HER is shown in Figure 6. Among the investigated ions, the single factor included Mg2+, Cu2+, and Na+ (Figure 6A,B) and the combined factor included Mg2+/Cu2+, Cu2+/Na+, Na+/NH4 +, Na+/K+, and NH4 +/K+ (Figure 6C–E) as the most influential factors for HER. The responding patterns of HER towards different kinds of ions appeared to be appreciably different. These effects can be broadly classified as counter-effective and synergistic. For instance, for the counter-effective impact, the binary Mg2+/Cu2+ belongs to this category, as does the binary NH4 +/K+ (Figure 6A,E). For the synergistic effect, the binary Cu2+/Na+, Na+/NH4 +, and Na+/K+ fall into this category (Figure 6B–D). These different ions will act as essential nutritious elements during metabolism at different stages of the growth of microbes [70–72]. For the growth pattern of microbes, there will normally be lagging, exponential, stationary, and death phases [73–75]. After inoculation, the microbes will experience a lagging phase with different duration [76,77]. The length of the lagging phase depends on many factors, such as the harshness of cultivation media, which contains lignocellulosic precursors and high levels of salt concentration [78–80].


**Table 4.** ANOVA analysis for the effect of ion concentration upon HER.

In this table, r2 = 0.94, adjusted r2 = 0.93, predicted r<sup>2</sup> = 0.93, and adequate precision (AP) = 15.

The strategies for how to improve and shorten the length of the lagging phase will contribute to the improvement of the duration of the lagging phase [81]. For microbes to initiate their metabolism, elements such as Mg2+, Na+, NH4 +, and K+ are essential [82–84]. These elements usually act as the major components of active centers in many enzymes [85–87]. Ensuring a sufficient amount of these necessary elements will facilitate the smooth and fast transition from the lagging phase to the growth phase [88–90]. It is commonly accepted that BioH2 generation will occur mainly in the exponential and stationary phases [91,92]. Clearly, these investigated literature reports provide useful guidance for the levels of these necessary ion elements in the cultivation media. More importantly, through statistical analysis from our developed ANN-RSM algorithm, the level of the response (the enhanced HER) for those inputs was underpinned. In addition, the order of significance for HER was also identified as the following: Na<sup>+</sup> > Mg2+ > Cu2+ > NH4 <sup>+</sup> > K+. From a holistic point of view, all the steps involved in BioH2 generation metabolism could be targeted as steps to enhance BioH2 generation (HY and HER). Two major metabolic pathways, namely, butyrate and acetate, are mainly associated with the activities of hydrogenase and the generation of H2 during dark fermentation [93–95]. From a stoichiometric perspective, the metabolic route towards acetate generates two times that of butyrate pathways [96,97]. From a process intensification point of view, the facilitation of the metabolic pathway towards an acetate pathway is favorable. From our statistical analysis, of all the investigated ions among the literature reports, the yield of BioH2 generation for these chemical additions of ions is not significant, suggesting that the enhancement of BioH2 generation by simple chemical additions of ions might be ineffective at further improving the ceiling value of BioH2 generation yield. For the sake of skewing the delicate balance between butyrate and acetate pathways, the combination of other chemical additions such as acti-

vated carbon, biochars, or porous adsorbents will be more effective in enhancing BioH2 generation [98–100].

**Figure 6.** ANNs-RSM analysis of statically significant ion concentrations for HER. (**A**) Mg2+/Cu2+ nanoparticle concentration versus HER, (**B**) Na+/Mg2+ nanoparticle concentration versus HER, (**C**) Na+/NH4 <sup>+</sup> nanoparticle concentration versus HER, (**D**) Na+/K+ nanoparticle concentration versus HER, (**E**) NH4 +/K+ nanoparticle concentration versus HER.

#### **6. Conclusions**

The statistical significance of these different NPs and ion additions were rigorously and quantitatively analyzed through a well-developed ANNs-RSM algorithm. As a result, this work provided effective guidance for the size optimization of NP additions and concentration regulation of ion additives in practice. For Fe-based NPs and ions, both the size of NPs and their corresponding concentration are statistically significant to HY. For HER, it was found that the combined effect of NP size and concentration is insignificant to HER. For Ni-based NPs and ions, neither size nor concentration is statistically significant to HY and HER, respectively. The variation in the size of NPs for the enhancement of

HY and HER behaved differently. The smaller (less than 42 nm) were found to definitely improve HY. Simultaneously, for HER, most reported literature indicated that manipulating the size of NPs is ineffective. It was found that variations in the investigated ions only statistically influenced HER, but not HY. This discovery suggests very pivotal guidance for process intensification for BioH2 generation. Using the constructed algorithm, the level of responses (enhanced HER) towards inputs (other ion additions) was underpinned, and the order of significance towards HER was also identified as the following: Na+ > Mg2+ > Cu2+ > NH4 <sup>+</sup> > K+. However, the number of relevant literature reports is currently limited; with the support of more experimental data, the results predicted by the ANNs-RSM algorithm will be more credible.

**Supplementary Materials:** The following are available online at https://www.mdpi.com/article/10 .3390/en14185916/s1, Figure S1: Schematic diagram of methodology: (A) The procedures flowchart, (B) ANNs construction: feed forward three layers networks; Table S1: Ion comparison upon BioH2 generation—refers to all data missing as, for convenience of calculation, the missing value was replaced by the averaged value during the artificial neuron network learning process; Table S2: ANOVA analysis for the effect of ion concentration on HY, where r2 = 0.94, adjusted r2 = 0.93, predicted r2 = 0.93, and adequate precision (AP) = 15.

**Author Contributions:** Drafting and data collection, Y.L., J.L. and Y.W.; paper writing and data collection, H.H.; proofreading, S.Y.; programming and modelling, J.H.; supervision, H.J. and T.C.; drafting, G.Y.; funding acquisition and project management, Y.S. All authors have read and agreed to the published version of the manuscript.

**Funding:** This work was supported by the Key Laboratory of Carbonaceous Wastes Processing and Process Intensification of Zhejiang Province (2020E10018), the Qianjiang Talent Scheme (QJD1803014), the Ningbo Science and Technology Innovation 2025 Key Project (2020Z100), and the Ningbo Municipal Commonweal Key Program (2019C10033 & 2019C10104), UNNC FoSE Researchers Grant 2020 (I01210100011).

**Institutional Review Board Statement:** Not applicable.

**Informed Consent Statement:** Not applicable.

**Acknowledgments:** The authors would like to express their sincere appreciation for the critical and insightful comments raised by anonymous reviewers, which significantly improved the quality of this work.

**Conflicts of Interest:** The authors declare no conflict of interest.
