The extraction of α-CD was performed using the Solvent process described by Li et al. [15] where the α-CGTase breaks and rearranges starch obtained from yams and cassava to α-CD units. In brief, the recovered starch (5% w/v) was suspended in 100 mL of distilled water. The slurry was heated to 92˚C to completely gelatinize the starch. The gelatinized starch solution was then cooled to 50˚C and the pH adjusted to 7 - 8 using sodium phosphate dibasic and NaOH. The

enzyme, α-CGTase (100 µL) and a complexing agent, butan-1-ol or octan-1-ol (5% v/v) were added after 2 h duration of adding the former. The reaction was left to run at 50˚C with continuous stirring at 200 rpm. The reaction content was analyzed after 4 and 7 days. The enzyme was then deactivated by 1 M HCl and the pH of the mixture was adjusted to 7 using 1 M NaOH. The insoluble CD-butanol complex was obtained after centrifuge, and it was resuspended in water. The complex was broken down to free the CD by distilling off the butanol. The α-CD was crystallized at 5˚C and analyzed by UV-vis spectroscopy using PerkinElmer Lambda 365 UV/VIS Spectrophotometer.

300 g of electronic scraps (CPU, RAM) were dissolved in a 500 mL sodium bromide/nitric acid mixture (NBN) or AR. The digest was characterized by atomic absorption spectroscopy (AAS) using AA500 from PG Instrument to determine the amount of gold present. The pH which was <2 forming HAuBr₄ was then adjusted to 4 - 6 by KOH (1 M) for optimum recovery by α-CD. The filtrate (50 mL) was mixed with crude α-CD (~0.1 M, 10 mL) forming a reddish-brown precipitate. The precipitate was dispersed in 50 mL of distilled water and then reduced by Na₂S₂O₃ (4 g) to elemental gold in a sludge form. For recovery of gold from the soil, 150 g was dissolved in 250 mL of NBN or AR. The recovered gold was analyzed by AAS. The reaction was repeated using pure commercially available α-CD (0.1 M) for comparison.

The first step was to extract starch from cassava and yams using a method described by Adikwu et al. [16] but with some modification. Blue-black colouration using iodine was used as a positive indicator of starch. α-CD was synthesized by the action of the enzyme α-CGTase on the amylose component of the starch. The amylose/amylopectin content was determined where 34.38% and 41.25% amylose was present in cassava and yam starch, respectively.

Gelatinized starch was refluxed with α-CGTase extracted from E. coli (≥90 U/mL activity). The reaction was monitored by UV-vis spectroscopy where both the pure reference standard and the synthesized α-CD (crude) had a ʎ_max of 220 nm (Figure 2). Yam starch yielded more α-CD than cassava starch as shown by the higher absorbance in the UV-vis spectrum (Figure 2) of the former which is

in agreement with the higher amylose (41.25% vs 34.38%) content in yams compared to cassava.

The synthesized crude α-CD was further checked with Fourier transform infrared (FTIR) using Shimadzu IRTracer-100 spectrometer for the presence of the hydroxyl functional groups (O-H bond absorption at 3288 cm⁻¹) and key fingerprint sections making the skeleton of α-CD indicated by C-H, C-O and C-O-C with the very close IR absorption bands in both pure and crude α-CD extracted (Figure 3).

The reaction was optimized for temperature, reaction time, nature of the complexing agent to recover α-CD from the solution phase, and the pH and nature of the medium used (Table 1). Cyclodextrins are highly soluble in water posing difficulty in obtaining a sufficient quantity in solid form. The quantification reported in Table 1 was for crude extract analyzed by UV-vis spectroscopy with reference to a calibration line of standard α-CD. Recovery of α-CD at pH 8.0 (35%) was high compared to pH 7.0 (29.2%) in sodium phosphate dibasic buffer with pH adjusted by aqueous sodium hydroxide (Table 1, entries 4 and 5). It is worth saying that distilled water (pH 6.8) was our solvent of choice with a yield of 84.9% of crude α-CD (Table 1, entry 1). It was observed the starch to α-CD conversion required long reaction times with a significant improvement from 31.1% to 35% after 4 and 7 days, respectively (Table 1, entries 1 and 2). Nevertheless, sufficient CD recovery is possible from starch sources with high amylose content as observed with yam starch (33.1%) after 4 days compared to the 31.1% synthesized α-CD from cassava starch in the same duration (Table 1, entries 2 and 3). Octanol was a better complexing agent recovering more α-CD than butanol (54.5% vs 35%) (Table 1, entries 1 and 6). This is in agreement with what has been reported in the literature [17]. A recovery of 63.2% crude α-CD was achieved using 20 g of starch (Table 1, entry 7) promising large-scale production.

¹Unless stated, all reactions were performed using α-CGTase (100 μL, ≥90 U/mL activity), 0.5 g of starch in 10 mL distilled water at 50˚C with continuous shaking at 200 rpm for 96 - 168 h. ^aReactions were performed in sodium phosphate dibasic. ^b20 g of starch in 100 mL of distilled water using 900 μL of the enzyme.

It is worthy to mention that performing the reaction at 60˚C gave no significant improvement in the yield of α-CD (Figure 3). A higher temperature could not be used to prevent denaturation of the enzyme and the formation of undesired by-products evident from the rough baseline. Further, yams were not used as the starting material for the synthesis of α-CD despite high amylose content and the shorter reaction time, because of the low starch (11%) obtained compared to 20% starch by mass using cassava. In addition, crude α-CD from yams was pale lilac which required further treatment before spectroscopic analysis.

With the sufficient recovery of α-CD (>60%), its ability to coordinate with gold was investigated. All reactions were performed in the same pot without the need to first isolate α-CD. The pH of the mixture was adjusted to 4 - 6 by sodium hydroxide to obtain optimal conditions for gold-CD coordination. Crude α-CD was added to the dissolved metal ions and the Au ions were precipitated as a reddish-brown α-CD-Au complex. The precipitate was recovered by filtration and reduced by sodium thiosulphate to obtain the gold. The presence of gold was confirmed qualitatively by the dark colouration using stannous chloride and was quantified by AAS. NBN proved better than AR in digesting electronic waste. This was evident from the higher concentration of gold in NBN (Figure 4). On the contrary, the amount of Au in the NBN digest of the gold-containing soil sample was lower than in AR (13.33 ppm vs 18.33 ppm). This is probably due to fewer metals in the soil allowing mostly the presence of gold in the AR digested mixture. Also, the numerous contents in electronic waste both metallic and plastic could have led to side reactions that prevented free gold ions in digested mixture analyzed.

The optimal pH for efficient α-CD-Au coordination was also investigated (Table 2). The per cent recoveries of Au by crude α-CD were 22.9%, 19.1% and 19.7% at pH 4, 5 and 6, respectively (Table 2, entries 1 - 3). Although there is no clear correlation of recoveries with the rising pH using crude α-CD, however with a 2-fold higher concentration of pure α-CD which recovered 78.8%, 37.5% and 64.2% of Au at pH 4, 5 and 6, respectively (Figure 5) indicating the reaction is favoured by low pH. Nevertheless, 70.8% Au recovery was achieved by crude α-CD from the NBN-soil digest of the sample obtained from an artisanal gold mine (Table 2, entry 4). For gold recovery from the soil sample, crude α-CD almost matches purified α-CD (79.2%).

¹Unless stated all reactions were performed using crude α-CD (0.1 M), electronic-waste (CPU and RAM chips) stock solution containing 87.33 ppm of gold in NaBr/HNO₃ (pH 4 - 6). ^aSoil containing 13.333 ppm of Au.