乙醇高值转化制正丁醇的机器学习研究

doi:10.6023/A26040110

摘要/Abstract

将生物乙醇转化为正丁醇是实现可再生能源长期储存和高值化利用的重要路径. 从庞大的合成空间中获得高效正丁醇合成催化剂的传统试错方法耗时耗力, 效率低, 且难以涵盖大量的材料组合. 基于此, 本研究引入机器学习工作方法, 旨在探索以提高正丁醇收率为目标的催化剂筛选和设计的数据驱动研究范式. 本研究收集了2000年以来的乙醇催化转化制正丁醇反应的非均相催化研究文献数据, 构建了一个包含催化剂组成、制备方法、结构特性、反应条件及其相应催化性能的乙醇转化制正丁醇的催化剂数据集. 基于该数据集, 评估了九种重要的回归算法, 其中, 极端梯度提升回归(XGBR, eXtreme Gradient Boosting)模型的训练集的决定系数(R²)为0.986, 测试集为0.897, 展现出超越其他模型的预测性能和泛化能力. 结合贝叶斯优化算法, 在正丁醇合成催化剂参数空间中高效筛选出5组高性能候选催化剂, 为乙醇催化转化制正丁醇反应催化剂的高效筛选与理性设计提供了新的思路, 为乙醇高值高效利用开辟了数据驱动新途径.

关键词: 乙醇催化转化, 正丁醇, 机器学习, 催化剂筛选, 贝叶斯优化

The conversion of bioethanol to n-butanol represents a crucial pathway for the long-term storage and high-value utilization of renewable energy. Traditional trial-and-error approaches to obtain efficient catalysts for n-butanol synthesis from vast synthetic spaces are time-consuming, labor-intensive, and inefficient, while being incapable of covering extensive material combinations. This study introduces machine learning methodologies to explore a data-driven research paradigm for catalyst screening and design aimed at improving n-butanol yield. We collected literature data on heterogeneous catalysis for ethanol-to-n-butanol conversion since 2000, establishing a comprehensive catalyst dataset encompassing catalyst composition, preparation methods, structural characteristics, reaction conditions, and corresponding catalytic performance. Nine regression algorithms were evaluated based on this dataset, with the eXtreme Gradient Boosting (XGBR) model achieving coefficient of determination (R²) values of 0.986 and 0.897 for the training and test sets, respectively, demonstrating superior predictive performance and generalization capability compared to other models. Coupled with Bayesian optimization, five high-performance candidate catalysts were efficiently identified within the n-butanol synthesis catalyst parameter space. This work provides novel insights for the efficient screening and rational design of catalysts for ethanol-to-n-butanol conversion and opens a data-driven avenue for the high-value and efficient utilization of ethanol.

Key words: ethanol catalytic conversion, n-butanol, machine learning, catalyst screening, Bayesian optimization

引用此文

张宇明, 孙世涛, 周诗嘉, 李珺卿, 贺雷, 王东琪. 乙醇高值转化制正丁醇的机器学习研究[J]. 化学学报, 2026, 84(9): 1464-1474.

Yuming Zhang, Shitao Sun, Shijia Zhou, Junqing Li, Lei He, Dongqi Wang. Value-added Transformation of Ethanol to n-Butanol: a Machine Learning Study[J]. Acta Chimica Sinica, 2026, 84(9): 1464-1474.

导出引用 EndNote|Reference Manager|ProCite|BibTeX|RefWorks

分享此文

图/表 11

图1. 数据集中催化剂组分分布: (a)主要活性金属; (b)载体种类

Figure 1. Distribution of catalyst components in the dataset: (a) major active metals; (b) types of support materials

图2. 由乙醇转化率、正丁醇选择性和正丁醇收率组成的散点图(收率随数值大小由不同颜色表示)

Figure 2. A scatter plot of ethanol conversion rate, n-butanol selectivity, and n-butanol yield (where yield values are represented by different colors according to their magnitude)

图3. 数据集中催化剂性能的分布直方图

Figure 3. Histogram of catalyst performance distribution in the dataset

图4. 预测特定特征缺失值的ETR模型: (a)比表面积; (b)孔容; (c)碱酸比; (d)重时空速

Figure 4. ETR models for predicting missing values of specific characteristics: (a) specific surface area; (b) pore volume; (c) base-to-acid ratio; (d) gravimetric space velocity

图5. 所有输入特征之间的相关性分析: (a)皮尔森相关性; (b)斯皮尔曼相关性

Figure 5. Correlation analysis among all input features: (a) Pearson correlation; (b) Spearman correlation

图6. 筛选后特征之间的相关性分析: (a)皮尔森相关性; (b)斯皮尔曼相关性

Figure 6. Correlation analysis between screened features: (a) Pearson correlation; (b) Spearman correlation

ML model	R²-train	R²-test	RMSE(%)-train	RMSE(%)-test
LR	0.350	0.373	7.39	7.62
RR	0.350	0.372	7.39	7.62
LLR	0.311	0.307	7.61	8.01
KNN	0.856	0.753	3.48	4.78
DTR	0.851	0.707	3.54	5.20
RFR	0.942	0.854	2.21	3.68
ETR	0.987	0.868	1.04	3.50
GBDT	0.975	0.869	1.46	3.48
XGBR	0.986	0.897	1.07	3.09

表1. 不同机器学习算法回归结果对比

Table 1. Comparison of regression results for different machine learning algorithms

图7. 不同机器学习模型的验证结果. LR: 线性回归; RR: 岭回归; LLR: 套索回归; KNN: K近邻回归; DTR: 决策树回归; RFR: 随机森林回归; ETR: 极端随机树回归; GBDT: 梯度提升回归; XGBR: 极端梯度提升回归

Figure 7. Validation results of different machine learning models. LR: Linear Regression; RR: Ridge Regression; LLR: Lasso Regression; KNN: K-Nearest Neighbor; DTR: Decision Tree Regression; RFR: Random Forest Regression; ETR: Extremely-randomized Trees Regression; GBDT: Gradient Boosting Decision Tree; XGBR: eXtreme Gradient Boosting Regression

图8. 不同机器学习模型的预测准确度评价: (a)模型的决定系数, R2; (b)模型的均方根误差, RMSE (%)

Figure 8. Evaluation of prediction accuracy for different machine learning models: (a) Model coefficient of determination, R²; (b) Root mean square error, RMSE (%)

图9. 通过SHAP分析得出的特征重要性(由上至下, 重要性依次减弱)

Figure 9. The feature importance derived from SHAP analysis (from top to bottom, importance decreases progressively)

No.	Metal	Support	Reaction temperature/℃	Butanol yield/%
1	Au	MgAlO_x	334	57.1
2	Ag	NaX	322	53.9
3	Cu	NC	344	51.1
4	Ca	NC	332	48.7
5	Be	NC	327	46.9

表2. 高性能正丁醇合成候选催化剂的模型预测结果

Table 2. Model prediction results for candidate catalysts for high-performance n-butanol synthesis

参考文献 60

[1]	Li H. Y.; Zheng Y. M.; Ji Y. L.; Yu M.; Shen Y. T.; Zhang S. R.; Liu Y. Y.; Wang H. Y.; Chen Z. W.; Dai X. D. Nat. Gas Oil 2025, 43, 129 (in Chinese).
	(李洪言, 郑一鸣, 季银利, 于淼, 沈奕天, 张胜然, 刘永煜, 王海洋, 陈子文, 代晓东, 天然气与石油, 2025, 43, 129.)
[2]	Wang J.; Kong L. K.; Wu Y. R.; Gu M. L.; Chen X. Y. Chem. Ind. 2025, 43, 33 (in Chinese). doi: 10.1021/ie50493a017
	(王敬, 孔令凯, 吴亚冉, 顾明龙, 陈翔宇, 化学工业, 2025, 43, 33.)
[3]	Li W.; Ghosh A.; Bbosa D.; Brown R.; Wright M. M. ACS Sustain. Chem. Eng. 2018, 6, 16515. doi: 10.1021/acssuschemeng.8b03622
[4]	Yang Z.; Liu S. M.; Xing L. Chem. Ind. 2025, 43, 1 (in Chinese).
	(杨铮, 刘思明, 邢磊, 化学工业, 2025, 43, 1.)
[5]	He L.; Zhou B.; Sun D.; Li W.; Lv W.; Wang J.; Liang Y.; Lu A. ACS Catal. 2023, 13, 11291. doi: 10.1021/acscatal.3c01481
[6]	Xie S.; Li Z.; Luo S.; Zhang W. Renew. Sust. Energ. Rev. 2024, 192, 114240. doi: 10.1016/j.rser.2023.114240
[7]	Eagan N. M.; Kumbhalkar M. D.; Buchanan J. S.; Dumesic J. A.; Huber G. W. Nat. Rev. Chem. 2019, 3, 223. doi: 10.1038/s41570-019-0084-4
[8]	Sargent M. J.; Slater N. K. H.; Dennis J. S.; Smith G. J.; Pearlman P. S. Sep. Purif. Technol. 2022, 288, 120655. doi: 10.1016/j.seppur.2022.120655
[9]	Wei M.; Zhou Q.; Liu J. Energy Sci. Eng. 2025, 13, 2695. doi: 10.1002/ese3.v13.6
[10]	Zhang C.; Si Z.; Chen B.; Chen C.; Chen H.; Ren W.; Cheng S.; Li S.; Cai D.; Qin P. J. Clean Prod. 2021, 285, 124902. doi: 10.1016/j.jclepro.2020.124902
[11]	Pasel J.; Haeusler J.; Schmitt D.; Valencia H.; Mayer J.; Peters R. Appl.Catal. B-Environ. Energy. 2023, 324, 122286.
[12]	Zhang S.; Chen J.; Wei B.; Zhou H.; Hua K.; Liu X.; Wang H.; Sun Y. J. Am. Chem. Soc. 2024, 146, 6037. doi: 10.1021/jacs.3c13092
[13]	Muck J.; Kocik J.; Hajek M.; Tisler Z.; Frolich K.; Kasparek, A. Appl. Catal. A-Gen. 2021, 626, 118380.
[14]	Xu X.; Zhang J.; Dong B.; Li F. Inorg. Chem. 2024, 63, 3843. doi: 10.1021/acs.inorgchem.3c04142
[15]	Chakraborty S.; Piszel P. E.; Hayes C. E.; Baker R. T.; Jones W. D. J. Am. Chem. Soc. 2015, 137, 14264. doi: 10.1021/jacs.5b10257 pmid: 26526779
[16]	Koda K.; Matsu-Ura T.; Obora Y.; Ishii Y. Chem. Lett. 2009, 38, 838. doi: 10.1246/cl.2009.838
[17]	Dowson G. R. M.; Haddow M. F.; Lee J.; Wingad R. L.; Wass D. F. Angew. Chem.-Int. Ed. 2013, 52, 9005. doi: 10.1002/anie.v52.34
[18]	Wingad R. L.; Gates P. J.; Street S. T. G.; Wass D. F. ACS Catal. 2015, 5, 5822. doi: 10.1021/acscatal.5b01327
[19]	Fu S.; Shao Z.; Wang Y.; Liu Q. J. Am. Chem. Soc. 2017, 139, 11941. doi: 10.1021/jacs.7b05939
[20]	Neumann C. N.; Rozeveld S. J.; Yu M.; Rieth A. J.; Dinca M. J. Am. Chem. Soc. 2019, 141, 17477. doi: 10.1021/jacs.9b08968
[21]	Kulkarni N. V.; Brennessel W. W.; Jones W. D. ACS Catal. 2018, 8, 997. doi: 10.1021/acscatal.7b03653
[22]	Di Cosimo J. I.; Apestegu, A C. R.; Ginés M. J. L.; Iglesia E. J. Catal. 2000, 190, 261. doi: 10.1006/jcat.1999.2734
[23]	Carvalho D. L.; de Avillez R. R.; Rodrigues M. T.; Borges L. E. P.; Appel L. G. Appl. Catal. A-Gen. 2012, 415, 96.
[24]	Wang Q.; Zhou B.; Weng X.; Lv S.; Schueth F.; Lu A. Chem. Commun. 2019, 55, 10420. doi: 10.1039/C9CC05454E
[25]	Tsuchida T.; Kubo J.; Yoshioka T.; Sakuma S.; Takeguchi T.; Ueda W. J. Catal. 2008, 259, 183. doi: 10.1016/j.jcat.2008.08.005
[26]	Yuan B.; Zhang J.; An Z.; Zhu Y.; Shu X.; Song H.; Xiang X.; Wang W.; Jing Y.; Zheng L.; He J. Appl. Catal. B-Environ. Energy. 2022, 309, 121271.
[27]	Riittonen T.; Toukoniitty E.; Madnani D. K.; Leino A.; Kordas K.; Szabo M.; Sapi A.; Arve K.; Warna J.; Mikkola J. Catalysts 2012, 2, 68. doi: 10.3390/catal2010068
[28]	Zhang J.; Shi K.; An Z.; Zhu Y.; Shu X.; Song H.; Xiang X.; He J. Ind. Eng. Chem. Res. 2020, 59, 3342. doi: 10.1021/acs.iecr.9b06778
[29]	Pang J.; Zheng M.; He L.; Li L.; Pan X.; Wang A.; Wang X.; Zhang T. J. Catal. 2016, 344, 184. doi: 10.1016/j.jcat.2016.08.024
[30]	Wu X.; Cai X.; Zhang Q.; Bi P.; Meng Q.; Pi Y.; Wang T. ACS Sustain. Chem. Eng. 2021, 9, 11269. doi: 10.1021/acssuschemeng.1c04305
[31]	Lv W.; He L.; Li W.; Zhou B.; Lv S.; Lu A. Green Chem. 2023, 25, 2653. doi: 10.1039/D3GC00078H
[32]	Wang Z.; Yin M.; Pang J.; Li X.; Xing Y.; Su Y.; Liu S.; Liu X.; Wu P.; Zheng M.; Zhang T. J. Energy Chem. 2022, 72, 306. doi: 10.1016/j.jechem.2022.04.049
[33]	Quesada J.; Faba L.; Diaz E.; Ordonez S. ChemCatChem 2018, 10, 3583. doi: 10.1002/cctc.v10.16
[34]	Lin X.; Fei X.; Chen D.; Qi Y.; Xu Q.; Liu Y.; Zhang Q.; Li S.; Wang T.; Qin Y.; Qiu X. ACS Catal. 2022, 12, 11573. doi: 10.1021/acscatal.2c02440
[35]	Jordison T. L.; Lira C. T.; Miller D. J. Ind. Eng. Chem. Res. 2015, 54, 10991. doi: 10.1021/acs.iecr.5b02409
[36]	Riittonen T.; Eranen K.; Maki-Arvela P.; Shchukarev A.; Rautio A.; Kordas K.; Kumar N.; Salmi T.; Mikkola J. Renew. Energy. 2015, 74, 369. doi: 10.1016/j.renene.2014.08.052
[37]	He J.; Li X.; Kou J.; Tao T.; Shen X.; Jiang D.; Lin L.; Li X. Catal. Sci. Technol. 2023, 13, 170. doi: 10.1039/D2CY01554D
[38]	Earley J. H.; Bourne R. A.; Watson M. J.; Poliakoff M. Green Chem. 2015, 17, 3018. doi: 10.1039/C4GC00219A
[39]	Wang Z.; Pang J.; Song L.; Li X.; Yuan Q.; Li X.; Liu S.; Zheng M. Ind. Eng. Chem. Res. 2020, 59, 22057. doi: 10.1021/acs.iecr.0c04649
[40]	Liu H.; Zheng T.; Hui T.; Zhang R.; Meng X.; Zhu W.; Liu H.; Liu Z. Chem. Eng. J. 2024, 481.
[41]	Lu B.; Ma S.; Liang S.; Wang Z.; Liu Y.; Mao S.; Ban H.; Wang L.; Wang Y. ACS Catal. 2023, 13, 4866. doi: 10.1021/acscatal.2c06213
[42]	Vermoortele F.; Ameloot R.; Vimont A.; Serre C.; De Vos D. Chem. Commun. 2011, 47, 1521. doi: 10.1039/C0CC03038D
[43]	Jiang D.; Fang G.; Tong Y.; Wu X.; Wang Y.; Hong D.; Leng W.; Liang Z.; Tu P.; Liu L.; Xu K.; Ni J.; Li X. ACS Catal. 2018, 8, 11973. doi: 10.1021/acscatal.8b04014
[44]	Zhou J.; He Y.; Xue B.; Cheng Y.; Zhou D.; Wang D.; He Y.; Guan W.; Fang K.; Zhang L.; Ni J.; Li X. Sustain. Energy Fuels. 2021, 5, 4628. doi: 10.1039/D1SE01002F
[45]	Wu X.; Fang G.; Liang Z.; Leng W.; Xu K.; Jiang D.; Ni J.; Li X. Catal. Commun. 2017, 100, 15. doi: 10.1016/j.catcom.2017.06.016
[46]	Jiang D.; Wu X.; Mao J.; Ni J.; Li X. Chem. Commun. 2016, 52, 13749. doi: 10.1039/C6CC05860D
[47]	Hu J. Y.; Yang J. Y.; Hu S. L.; Liu J. X.; Li W. X. Sci. Sin. Chim. 2025, 55, 1704 (in Chinese). doi: 10.1360/SSC-2025-0066
	(胡建钰, 杨峻毅, 胡素磊, 刘进勋, 李微雪, 中国科学(化学), 2025, 55, 1704.)
[48]	Li J. Q.; Song Q. X.; Liu Z. Y.; Wang D. Q. Acta Chim. Sinica 2024, 82, 387 (in Chinese). doi: 10.6023/A23100473
	(李珺卿, 宋千禧, 刘子义, 王东琪, 化学学报, 2024, 82, 387.) doi: 10.6023/A23100473
[49]	Dong Y.; Ran M.; Zhang X.; Lin S.; Zhao H.; Yang Y.; Liu S.; Zheng C.; Gao X. ACS EST Eng. 2024, 4, 1312. doi: 10.1021/acsestengg.3c00617
[50]	Toyao T.; Maeno Z.; Takakusagi S.; Kamachi T.; Takigawa I.; Shimizu K. ACS Catal. 2020, 10, 2260. doi: 10.1021/acscatal.9b04186
[51]	Takahashi K.; Ohyama J.; Nishimura S.; Fujima J.; Takahashi L.; Uno T.; Taniike T. Chem. Commun. 2023, 59, 2222. doi: 10.1039/D2CC05938J
[52]	Olu-Ajayi R.; Alaka H.; Sulaimon I.; Sunmola F.; Ajayi S. J. Build. Eng. 2022, 45, 103406.
[53]	Jayakumar T. P.; Babu S. P. S.; Nguyen T. N.; Le S. D.; Manchan R. P.; Phulkerd P.; Chammingkwan P.; Taniike T. Appl. Catal. A-Gen. 2023, 666, 119427. doi: 10.1016/j.apcata.2023.119427
[54]	Tripathi K.; Gupta V.; Awasthi V.; Pant K. K.; Upadhyayula S. Adv. Sustain. Syst. 2023, 7, 2200416. doi: 10.1002/adsu.v7.3
[55]	Zhou Q.; Lu S.; Wu Y.; Wang J. J. Phys. Chem. Lett. 2020, 11, 3920. doi: 10.1021/acs.jpclett.0c00665
[56]	Dong Y.; Zhang Y.; Ran M.; Zhang X.; Liu S.; Yang Y.; Hu W.; Zheng C.; Gao X. J. Mater. Chem. A 2021, 9, 23850. doi: 10.1039/D1TA06772A
[57]	Li S.; Zhang Y.; Hu Y.; Wang B.; Sun S.; Yang X.; He H. J. Materiomics 2021, 7, 1029.
[58]	Li W.; Dong Y.; Ran M.; Lin S.; Liu P.; Song H.; Yi J.; Zhu C.; Qi Z.; Zheng C.; Zhang X.; Gao X. J. Energy Chem. 2025, 104, 372. doi: 10.1016/j.jechem.2024.12.056
[59]	Lundberg S. M.; Erion G.; Chen H.; Degrave A.; Prutkin J. M.; Nair B.; Katz R.; Himmelfarb J.; Bansal N.; Lee S. Nat. Mach. Intell. 2020, 2, 56. doi: 10.1038/s42256-019-0138-9 pmid: 32607472
[60]	Ahmad M. W.; Reynolds J.; Rezgui Y. J. Clean Prod. 2018, 203, 810. doi: 10.1016/j.jclepro.2018.08.207