3.1. Optimization glycolysis process
The glycolysis results obtained with each type of PET waste are submitted to a Taguchi methodology to determine the levels that maximize the BHET production. In a Taguchi design, the mean is the average response for each combination of control factor levels. So, the means provide an estimate of BHET yield at each factor level to know how the four control factors affect the yield of BHET. Delta is the difference between the highest and lowest average response values for each factor. Table 4 shows the response for the means, where the ranks are based on Delta values; Rank 1 to the highest Delta value, Rank 2 to the second highest, and so on, to indicate the relative effect of each factor on the response. Figure 1 illustrates the influence of the conditions on the yield of BHET.
Table 4
Table response for the glycolysis PET waste means
Level | T [ºC] | t [min] | EG [w/w] | ZnAc2 [%w/w] | Virgin PET |
1 | 76.59 | 78.79 | 67.62 | 77.11 |
2 | 77.19 | 76.27 | 76.54 | 76.04 |
3 | 77.02 | 75.74 | 86.64 | 77.65 |
Delta | 0.60 | 3.05 | 19.02 | 1.61 |
Rank | 4 | 2 | 1 | 3 |
Level | T [ºC] | t [min] | EG [w/w] | ZnAc2 [%w/w] | Multilayer PET |
1 | 66.83 | 68.15 | 57.21 | 66.61 |
2 | 66.76 | 65.35 | 65.10 | 65.07 |
3 | 65.34 | 65.42 | 76.62 | 67.25 |
Delta | 1.49 | 2.8 | 19.41 | 2.17 |
Rank | 4 | 2 | 1 | 3 |
Level | T [ºC] | t [min] | EG [w/w] | ZnAc2 [%w/w] | Highly colored PET |
1 | 74.26 | 76.00 | 64.43 | 74.35 |
2 | 74.29 | 73.28 | 73.77 | 72.84 |
3 | 74.04 | 73.31 | 84.40 | 75.41 |
Delta | 0.25 | 2.72 | 19.97 | 2.57 |
Rank | 4 | 2 | 1 | 3 |
Level | T [ºC] | t [min] | EG [w/w] | ZnAc2 [%w/w] | MSW PET |
1 | 73.92 | 74.79 | 64.36 | 72.96 |
2 | 73.26 | 72.64 | 72.69 | 72.30 |
3 | 72.82 | 72.56 | 82.94 | 74.74 |
Delta | 1.10 | 2.23 | 18.58 | 2.44 |
Rank | 4 | 3 | 1 | 2 |
According to the results obtained by Taguchi (Table 4 and Fig. 1), it is observed that the EG to PET mass ratio (EG/PET) is the most significant factor in all studied PET waste, with a great difference with respect to the others in the Delta value. The next most influential factor is time (in all cases except MSW PET), followed by the ZnAc2 to PET mass percentage ratio (ZnAc2/PET). Finally, the least influential factor is temperature, although these last three factors have a very similar and low influence (similar values of Delta) on BHET yield [47, 53]. Therefore, through these results it can be confirmed that the behaviour of the factors studied is the same with all types of PET waste, as well as virgin PET, according to other authors [36, 52].
Focusing on the EG to polymer mass ratio, the best level to achieve the highest BHET yield, is the one with the highest amount of solvent used (6 w/w) due to the yield of the reaction increase with the amount of EG as show the Taguchi results in Fig. 1. As expected and according to literature, the use of an excess of EG is the most important parameter due to the polycondensation step is reversible and in the presence of excess of EG, glycolysis of PET occurs to promote the reverse reaction to give BHET [35, 37, 51, 47, 55–56].
About other factor studied, Fig. 1 illustrates that an increase in time reaction decrease BHET yield. According to literature, as reaction time continues increasing, BHET starts to dimerise and polymerise into oligomers [44]. A similar observation was made by other authors, who found that after a long time of reaction, equilibrium is attained in the obtaining of BHET [39, 49, 57]. So, this is the reason why in this work the highest yield is obtained with the lowest time, due to with a suitable temperature and catalyst, 150 min time is enough time to obtain the maximum yield. On the other hand, if lower times had been used, an increase in BHET yield would have been observed with the increase in time [51].
The effect of using catalyst in the reaction is touchable which decreases the reaction time required [43]. Sometimes, with further increase in the amount of a catalyst, the yield of BHET decreased [55], but in other case increase [49, 56]. This is because this factor is very related and depends on the temperature and reaction time used. For example, ZnAc2 has the best efficiency when the temperature range is between 180 ºC to 195 ºC [43–44]. In this work, ZnAc2 has the best efficiency using the highest mass percentage ratio to PET (1%w/w) as it can be seen in Fig. 1 and which is according to literature [37, 56]. However, decreasing the mass percentage ratio of ZnAc2/PET to the lowest (0.2%w/w) only decreases the yield by 0.6% show a little influence of this factor on the glycolysis. So, due to economic and environmental reasons, it is not worth using the maximum level of catalyst.
Finally, temperature is the least influential factor on yield of BHET, having practically no influence compared to the others factors [20, 43]. Temperatures under 190 ºC produce an excess of sub-products and decrease the yield of BHET [48, 59] and higher temperatures over 208 ºC do not increase the yield of BHET [2, 43, 49]. This is in accordance with most of the reported works on glycolysis of PET which have been carried out at relatively low temperatures in the range of 190–200°C due to the yield of BHET at softening point temperatures is almost constant [47]. So, the best level chosen as optimum for this factor is the lowest (195 ºC) which agrees with literature [46].
According to Taguchi’s methodology, if the best levels of control factors are used (195 ºC, 150 min, 6 w/w EG/PET and 0.2%w/w ZnAc2/PET), a maximum BHET yield will be obtained whose value would be between 79–88% depending the type of PET used. As expected, virgin PET gives the highest BHET yield (88.33%), followed by highly colored PET (86.41%), MSW PET (84.61%) and multilayer PET (79.28%). In order to demonstrate the theoretical results of Taguchi’s study, a confirmatory experiment is carried out with the optimum conditions to verify the BHET yield expected. In this last confirmatory experiment, the results are close to the theoretical, with yields between 77–87%. These results have only a difference of 2% compared to theoretical as it can be seen in Fig. 2. More in detail, virgin PET also gives the highest BHET yield (86.82%), followed by highly colored PET decreasing yield by only 2.6% (84.52%), MSW PET decreasing by 4.8% (82.65%) and the last is multilayer PET with more difference, decreasing by 11.4% compared to the virgin PET (76.93%), due to has more impurities in its composition. The yield of BHET obtained is similar or even higher considering literature with similar conditions [36–38, 54–56, 58, 60–64]. It has been demonstrating that PET nature, colour and multilayer material has low influence on BHET yield, being the results of complex PET waste similar to virgin PET [36, 54].
3.2. Recovery and reuse of EG
One of the inconvenient of the glycolysis process is the excessive use of EG to obtain high yield of BHET. According to the study carried out in this work and the results obtained by Taguchi with the virgin PET, with an EG/PET mass ratio of 6 w/w, the BHET yield obtained is 88.33%. When the EG/ PET mass ratio decreases to 4 w/w, the BHET yield also reduce by 11.4% (78.23%), reaching a reduction of 21.5% (69.31%) in case of 2 w/w. This excessive use of EG has its drawback in the economy of the process. Increasing the EG/PET mass ratio from 4 to 6 means that the cost of producing 1kg of BHET increases by 28% [48]. Therefore, it is necessary to recover the remaining EG of the process and reuse it for consecutive reactions, in order to optimise the glycolysis of PET making it economical and more environmentally.
The filtrate obtained from BHET crystallization contain mainly unreacted EG and water, and catalyst but at lower levels. A distillation of this solution is carried out recovering the EG which vary from 70–85% of the initial amount. So, for comparison purposes, the EG used for each new glycolysis cycle is 70% of recovered EG and 30% of fresh EG. The aim is to determine the number of cycles conduct with this mixture of recovered and fresh EG and evaluate the evolution of the BHET yield and purity. These global cycles are repeated 10 times. Figure 3 illustrates the evolution of the BHET yield using recovered EG, along the 10 consecutive glycolysis reactions for the different PET wastes evaluated. It is found that yield is maintaining constant at around 71–81% depending the type of PET waste. In general, BHET yield decreased up to 6.6–7.4% from the glycolysis carried out with fresh EG to the glycolysis carried out with 10 times recovered EG and with the same behaviour all types of PET waste studied. Although this observation could lead to assume that recovered EG is as reactive as fresh EG, it has to be pointed out that these results may be affected by increased catalyst concentration during the consecutive reactions [58].
3.3. Analysis of products
In order to check the purity of the BHET recovered after different PET waste glycolysis, this was analysed by means of a number of analytical techniques, and the results were compared with those corresponding to a commercial BHET sample.
Figure 4 shows characteristic signals of BHET molecule through FTIR spectrum. Bands at 2867 and 2960 cm− 1 correspond to the stretching vibrations, asymmetric and symmetric, of the methyl and methylene groups (aliphatic CH2 group) which indicate a depolymerization reaction. Bands at 3434 and 3281 cm− 1 confirm the presence of O-H bond vibrations at chain terminations. The peak at 1132 cm− 1 confirms the presence of hydroxyl groups related to C-OH alcohol bonds, while the peaks between 1282 − 1020 cm− 1 are originated from the stretching vibrations of the C-O ester bonds. A transmission band at 1721 cm− 1 confirms the presence of ester carbonyl group (C = O). Bands at 1509 − 1404 and 720 cm− 1 correspond to the C = C bonds of the aromatic rings. All these results are according with literature [36, 38, 47, 64–65]. From all these observations it can be confirmed that the glycolysis products correspond with BHET, regardless of the type of PET waste and the use of recovered EG.
Figure 5 shows the DSC thermal analysis of the different BHET obtained from the glycolysis recorded by heating sample 250 ºC at 10 ºC·min− 1. All the samples show only one sharp endothermic peak at 110 ºC which agrees with the known melting point of commercial BHET (at 112 ºC) [36–37, 44–45, 47, 65–66]. Besides, it can be confirmed that the product is BHET with a high purity due to does not appear any other peak. For example, according to literature, BHET dimmers, trimers and oligomers appear at 151 ºC, 170 ºC and 210 ºC; and a peak in the region of 250 ºC is attributed to the melting of the remaining solid PET [44–45, 63]. So, DSC chromatogram also demonstrate high level of purity of the monomers produced by glycolysis process of different PET waste and by the addition of recovered EG to the process.