A method for predicting the quality of catalyst ligands for acetylene hydrochlorination based on deep neural networks
Through the deep neural network model combining catalyst ligands and reaction conditions, the LDNN model was constructed, which solved the problem of time-consuming and labor-consuming of traditional methods, and achieved efficient and accurate prediction of the advantages and disadvantages of catalyst ligands, and adapted to variable experimental conditions.
Patent Information
- Application Number
- CN202411670054.9
- Authority / Receiving Office
- CN · China
- Patent Type
- Patents(China)
- Current Assignee / Owner
- Filing Date
- 2024-11-21
- Publication Date
- 2025-08-22
- Estimated Expiration
- 2044-11-21
AI Technical Summary
The prior art consumes time and effort in screening highly efficient mercury-free acetylene hydrochlorination catalysts, and traditional methods fail to effectively consider the combined impact of catalyst ligands and experimental environments, resulting in insufficient prediction accuracy.
The deep neural network (DNN) model is used to combine the ECFP fingerprint and reaction conditions of the catalyst ligand to construct the LDNN model, and the model is trained through feature extraction and gradient enhancement methods to predict the conversion rate of the catalyst and guide the catalyst synthesis.
It improves the accuracy and efficiency of the prediction of the advantages and disadvantages of catalyst ligands, reduces the calculation cost, and can make high-precision predictions under small samples to adapt to unknown ligands and changing experimental conditions.
Smart Images
Figure CN119541732B_ABST
Abstract
Description
Technical Field
[0001] The present invention relates to a method for predicting the quality of acetylene hydrochlorination reaction catalyst ligands based on a deep neural network, and in particular to a method that combines the knowledge of catalysis and deep neural networks to design and implement an artificial intelligence-assisted screening of the quality of acetylene hydrochlorination reaction catalyst ligands, belonging to the technical field of materials informatics. Background Art
[0002] Polyvinyl chloride (PVC) is a crucial chemical raw material, primarily used in the production of plastics. PVC is in high demand and is primarily produced by the polymerization of vinyl chloride. Therefore, the synthesis of vinyl chloride is the most critical process in PVC production, primarily through the hydrochlorination of acetylene. Catalysts are essential in chemical production, and efficient acetylene hydrochlorination catalysts are particularly important. Traditionally, acetylene hydrochlorination has been performed using catalysts containing HgCl2. However, Hg, as a heavy metal, poses significant risks to humans and the environment. Given these environmental concerns, the search for equally efficient, clean, and pollution-free catalysts has become a key focus. Currently, research on mercury-free catalysts focuses on four main approaches: ruthenium-based, platinum-based, gold-based, and copper-based catalysts. This patent utilizes a machine learning model based on a deep neural network algorithm to predict high-performance acetylene hydrochlorination catalysts based on these four substrates.
[0003] The performance of a catalyst is usually affected by multiple factors, such as the size of the atom, the valence state of the element, and the structure of the catalyst. In the process of catalyst design and prediction, the traditional method is usually adopted, that is, to screen out efficient mercury-free hydrochlorination catalysts through experiments. In this process, a lot of time and financial costs are usually consumed. With the rapid development of current science and technology, artificial intelligence technology has been greatly promoted. Currently, artificial intelligence has been applied to various fields and is an excellent tool for assisting scientific research. This patent thinks of assisting our design and prediction process of efficient mercury-free hydrochlorination catalysts by training a machine learning model. Summary of the Invention
[0004] The present invention addresses the technical problem of proposing a deep neural network-based method for predicting the quality of catalyst ligands in acetylene hydrochlorination. This method uses a deep neural network (DNN) to model existing literature and experimental data, resulting in a new prediction model, named the LDNN model, which is used to predict the conversion of catalysts containing unknown ligands. The model's output guides catalyst synthesis. The method involves first reviewing various relevant literature and constructing a ligand database of chemical ligands derived from these studies. Next, a novel model is proposed that integrates a feature extractor and a DNN algorithm to predict the conversion of acetylene hydrochlorination catalysts containing unknown ligands. This model calculates the ECFP fingerprint of the relevant ligands based on their composite SMILES strings. This model then constructs a tensor matrix based on specific reaction conditions to provide a targeted prediction of the conversion, thereby improving prediction accuracy.
[0005] In order to solve the above technical problems, the present invention proposes a technical solution: a method for predicting the quality of acetylene hydrochlorination catalyst ligands based on a deep neural network, the method comprising the following steps:
[0006] (1) Collect and extract data from literature as a dataset for model training; catalyst ligands, reaction conditions and ligand properties are used to train the model;
[0007] (2) Extract the composite SMILES string of the ligand molecule and the reaction conditions and concatenate the tensor matrix; each atom and chemical bond in the composite SMILES string is a word, obtain the SMILES through public information, use ECFP (Extended Connectivity Fingerprint) technology to collect features, combine the reaction conditions, train iteratively, and use the Reul function;
[0008] (3) Input the tensor matrix into the model developed based on the deep neural network for training;
[0009] (4) Input the composite SMILES string of the unknown performance ligand and the reaction conditions to obtain the results of the ligand quality;
[0010] The reaction conditions and the ligand composite SMILES string are concatenated and a tensor matrix is constructed, which is input into the model as an input variable. The trained LDNN model is used to embed the ligand composite SMILES string into the tensor and output the quality of the catalyst.
[0011] Preferably, the input modules are divided into three categories, a buffer layer and a hidden layer are added to each module, and finally the BCEloss loss function is used for gradient improvement.
[0012] Preferably, the information extracted from the molecular library is input into the model and combined with the reaction conditions to predict the performance of the unknown ligand.
[0013] Preferably, the method comprises the following steps:
[0014] Step 1: Establish a data set and collect information on known catalyst ligands from the literature. The molecular structure of the ligand and the corresponding composite SMILES string, temperature, acetylene space velocity, central metal content, HCl ratio, and conversion rate are recorded in the database;
[0015] Step 2: Define the catalyst conversion rate improvement
[0016] Conversion rate improvement rate (C l ), specifically , C e is the conversion rate of the catalyst control sample without ligand under the same conditions, C r is the conversion rate of the catalyst with added ligand;
[0017] Step 3: Convert the molecular structure of the ligand into an ECFP fingerprint
[0018] The composite SMILES string of each ligand is converted into a 256-bit binary vector called an "ECFP fingerprint." For known ligand SMILES strings, a search is performed to map each functional group to a molecular fingerprint consisting of 0s and 1s from the RDKit open-source library, combining them into a 256-bit molecular fingerprint. Different sequences of 0s and 1s represent different functional groups, thus reflecting the influence of different ligand structures.
[0019] Step 4: Construct the feature tensor matrix
[0020] Each ligand collected has its own 256-bit molecular fingerprint. The fingerprint data of all ligands are integrated into a tensor matrix. Each row of the matrix represents a ligand and each column is a feature bit.
[0021] The experimental conditions temperature and acetylene space velocity are converted into vectors and added to the characteristic matrix;
[0022] Again, the experimental conditions of HCl ratio and central metal content are converted into vectors and added to the feature matrix;
[0023] Step 5: Build and tune the model architecture
[0024] Using a DNN algorithm, since the tensors containing airspeed and temperature differ significantly in magnitude, they are partitioned into separate input modules. A buffer layer is added as the first layer to reduce the differences between the different variables. For each of the three input modules, the corresponding function Reul is used to convert them into three vectors. These are then concatenated into a matrix and converted into a 1*1 tensor using the Reul function again. The BCE loss function is used to train the parameters of the corresponding function using gradient boosting backpropagation, and the conversion rate is predicted.
[0025] Step 6: Model training and evaluation
[0026] The simplified feature matrix group was input into the LDNN and training continued. The target variable was the conversion rate. Through model training, the training step size was adjusted to 0.001 and the number of training times was adjusted to 2000. The model gradually learned the relationship between ligand structure, experimental conditions and conversion rate.
[0027] At the same time, to prevent overfitting, the dropout function was used, discarding 12% of the data each time, and the optimal value of each training parameter was determined through trial and error. After training, the model was evaluated using the validation set. Finally, the model's prediction accuracy for different ligands was less than 5% with a probability of more than 80%.
[0028] Step 7: Predict the conversion rate of unknown ligand
[0029] After training, the model can be used to predict the conversion rate of a new ligand, ligand X. RDKit generates a 256-bit fingerprint for ligand X, which is then merged into a feature matrix along with the corresponding experimental conditions. After inputting the feature vector of X's experimental conditions, including temperature, HCl ratio, central metal content, and space velocity, the model outputs the predicted conversion improvement rate. This output can help determine whether ligand X has a good catalytic effect under the reaction conditions.
[0030] The method for predicting the quality of acetylene hydrochlorination catalyst ligands based on deep neural network is characterized in that the ligands are screened out.
[0031] As a catalyst for acetylene hydrochlorination.
[0032] A computer-readable storage medium having a computer program capable of running the method for predicting the quality of acetylene hydrochlorination catalyst ligands based on a deep neural network according to claim 1.
[0033] A device for predicting the quality of ligands for acetylene hydrochlorination catalysts based on a deep neural network, the device being equipped with the computer-readable storage medium of claim 6.
[0034] The method comprises the following steps:
[0035] (1) Construct a database of relevant chemical ligands for model training; (2) Extract the ligands and convert them into 256-bit ECFP fingerprints and separate them bit by bit, and then construct a tensor matrix based on the reaction conditions; (3) Input the tensor matrix as a preliminary feature vector into the DNN, where each column in the tensor matrix is a feature, set the importance attribute of each feature to the model performance, and obtain the importance parameter of each feature by preliminarily training the DNN model using database data. Arrange them in descending order and delete the features with relatively low importance in proportion to reduce the dimension and prevent overfitting; (4) Construct the environmental variables airspeed, temperature and HCL ratio into a tensor matrix of environmental variables, and merge it with the tensor matrix after dimensionality reduction to obtain a complete feature matrix; (5) Input the feature matrix into the DNN to train the model and evaluate its performance.
[0036] First, we collected a large amount of chemical ligand data by reviewing various relevant literature. Each ligand needed to include information such as its molecular structure, properties, and conversion rate in the acetylene hydrochlorination reaction. This information was obtained through literature research, laboratory data collection, and the publicly available chemical database PubChem.
[0037] Furthermore, the chemical ligands are characterized by ECFP (Extended Connectivity Fingerprint) technology. ECFP can effectively represent the topological information of the molecule and convert the chemical structure of the ligand into a fixed-length vector representation. The 256-bit ECFP fingerprint of each ligand is extracted using the chemical informatics tool RDKit. An iterative algorithm is used to extract the neighborhood information of atoms or groups in the molecule and construct a series of binary bit representations to generate a fixed-length molecular fingerprint. The extracted 64-bit ECFP fingerprint is converted into a tensor matrix. The ECFP fingerprint of each ligand is a 64-dimensional binary vector. When constructing the tensor matrix, the fingerprint of each ligand will become a row vector of a matrix. When there are multiple ligands, these row vectors will form a two-dimensional tensor, in which each row represents the fingerprint of a ligand.
[0038] Next, we train an LDNN model using the tensor matrix obtained in the previous step as the feature vector input and the conversion rate as the target output. As training progresses, the LDNN model uses gradient boosting backpropagation to optimize the corresponding hidden layer function at each iteration, adjusting the amplitude and number of training cycles. We also attempt to account for the influence of reaction conditions.
[0039] Furthermore, we selected experimental environmental variables closely related to the acetylene hydrochlorination reaction: space velocity, temperature, and HCl ratio. These variables all influence the catalyst's performance, making it essential to consider these factors in the model. We converted the environmental variables into a tensor matrix, with each entry's value as a column in the matrix, to construct a two-dimensional tensor of the environmental variables. The tensor matrix of the environmental variables was then merged with the reduced tensor matrix column by column to obtain a complete feature matrix.
[0040] Furthermore, the complete feature matrix obtained in the previous step is used as input, and the conversion rate is used as the target variable to input into the DNN model for training. During the training process, the model is trained using the gradient boosting method using the BCEloss function.
[0041] Beneficial effects:
[0042] This invention uses a deep-learning neural network (DNN)-based algorithm to train a model. The main advantage of DNNs is their ability to fit and predict data even with low fluxes and to perform well when the fluxes differ significantly. This type of model significantly reduces workload and offers high accuracy. We have used this algorithm to design and predict several high-performance mercury-free hydrochlorination catalysts, demonstrating high accuracy and catalytic performance.
[0043] Existing catalyst conversion rate prediction methods (such as DFT models) are typically based on solving the Schrödinger equation, which is complex and requires significant computing resources. Using a universal screening method for small AI models, we can blur and blackbox the intermediate processes, highlighting the input and output. For a trained model, predictions for a single ligand take only tens of milliseconds, significantly reducing computational costs.
[0044] Existing catalyst conversion rate prediction methods usually focus on data features from a single source, such as modeling based solely on the molecular structure information of the ligand or a single experimental environment variable. The flaw of such a model is that it ignores the interaction between the chemical ligand and the experimental environment variables and their combined impact on the catalytic reaction performance. This model combines the ECFP fingerprint of the ligand with the experimental environment variables to predict catalytic performance. In this way, the model not only takes into account the molecular structure of the ligand during the catalyst preparation process, but also can make dynamic predictions based on changes in experimental conditions, thereby more comprehensively capturing the factors affecting catalyst performance. This fusion of multiple features greatly improves the prediction accuracy of the model, especially when facing unknown ligands and arbitrarily changing experimental conditions.
[0045] Many traditional methods use simple molecular descriptors or features to represent ligands. These methods often fail to effectively capture the complexity of ligand structures, resulting in poor model performance. The present invention uses ECFP fingerprints to extract features from organic ligands and converts them into 256-bit binary vectors. By converting the ECFP fingerprint of each ligand into a tensor matrix and inputting it into the model, the present invention can efficiently express the complex information of the molecular structure. This tensor matrix form can capture higher-order and more hidden relationships in molecules, thereby improving the model's ability to represent ligands and its prediction accuracy. BRIEF DESCRIPTION OF THE DRAWINGS
[0046] The present invention will be further described below with reference to the accompanying drawings.
[0047] Figure 1 It is a schematic diagram of the present invention DETAILED DESCRIPTION
[0048] Example 1
[0049] The technical solution of the present invention is further explained below through embodiments in combination with the accompanying drawings.
[0050] Step 1: Create a dataset
[0051] First, collect information on known catalyst ligands from the literature. For example, consider a catalyst ligand A that achieves 70% conversion in acetylene hydrochlorination at 210°C and an HCl ratio of 1.10. Record this information (ligand molecular structure and corresponding composite SMILES string, temperature, acetylene space velocity, central metal content, HCl ratio, and conversion) in a database.
[0052] Step 2: Define the catalyst conversion rate improvement
[0053] In order to balance the impact of different batches of activated carbon from different manufacturers on the conversion rate, the conversion rate improvement rate (C l ) is used to describe, specifically C l = (C e is the conversion rate of the catalyst control sample without ligand under the same conditions, C r is the conversion rate of the catalyst with added ligand)
[0054] Step 3: Convert the molecular structure of the ligand into an ECFP fingerprint
[0055] To enable the model to "recognize" the ligand's structure, the composite SMILES string for each ligand is converted into a 256-bit binary vector called an "ECFP fingerprint." Using a search method for the SMILES string of a known ligand, each functional group is mapped to a molecular fingerprint consisting of 0s and 1s from the RDKit open-source library, resulting in a combined 256-bit molecular fingerprint. Different sequences of 0s and 1s represent different functional groups, thus reflecting the influence of different ligand structures.
[0056] For example, the composite SMILES string for ligand A is C=CBr. This string is retrieved from the RDKit library and converted into a 256-bit binary fingerprint: [1, 1, 0, 1, 0, 0, 1, ...]. Each bit in this string represents a detail of the molecular structure, such as the presence of specific chemical bonds or groups. This fingerprint serves as the ligand's "identity card" for subsequent model training.
[0057] Step 4: Construct the feature tensor matrix
[0058] Assume that each ligand collected has its own 256-bit molecular fingerprint. Combine the fingerprint data of all ligands into a tensor matrix. Each row of the matrix represents a ligand, and each column is a feature bit. For example, the first two rows can be like this:
[0059] | Ligand number | Position 1 | Position 2 | Position 3 | Position 4 | Position 5 | ... |
[0060] |----------|-----|-----|-----|-----|-----|-----|
[0061] | A | 1 | 1 | 0 | 0 | 1 | ... |
[0062] | B | 0 | 1 | 0 | 0 | 1 | ... |
[0063] Under this matrix structure, the molecular structural characteristics of all ligands are clear at a glance, making it convenient to input them into the model later.
[0064] Then, experimental conditions were added. The conversion rate of acetylene hydrochlorination is not only affected by the ligand structure but also closely related to experimental conditions (such as temperature and space velocity). To improve prediction accuracy, the experimental conditions (temperature and space velocity) were converted into vectors and added to the feature matrix.
[0065] Assume that the experimental temperature of ligand A is 210℃ and the space velocity is 340h -1 , these values can be encoded as additional feature columns, such as:
[0066] | Ligand number | Temperature | Space velocity |
[0067] |----------|------|----------|
[0068] | A | 210 | 340 |
[0069] | B | 250 | 680 |
[0070] In addition, due to the large difference in magnitude, in order to prevent other experimental conditions (HCl ratio and central metal content) that affect the conversion rate of acetylene hydrochlorination reaction from being masked by feature changes with large magnitude, these features are separated separately and composed into a new matrix for separate input.
[0071] Again, convert the experimental conditions such as HCl ratio and central metal content into vectors and add them to the feature matrix. Assuming that the HCl ratio of ligand A is 1.10 and the central metal content is 1%, encode these values as additional feature columns, such as:
[0072] | Ligand number | HCl ratio | Central metal content |
[0073] |----------|---------|---------------|
[0074] | A | 1.10 | 1% |
[0075] | B | 1.15 | 0.5% |
[0076] In this way, the feature matrix group not only contains the molecular structure information of the ligand, but also includes different experimental conditions, so that the model can take into account the differences in reaction performance caused by environmental changes.
[0077] Step 5: Build and tune the model architecture
[0078] After deciding to use the DNN algorithm, considering the significant difference in magnitude between the tensors containing airspeed and temperature, we partitioned them into separate input modules and added a buffer layer as the first layer to reduce the variance between the different variables. For each of the three input modules, we used the corresponding function Reul to convert them into three vectors. These were then concatenated into a matrix and converted into a 1*1 tensor using the Reul function again. The parameters of the corresponding functions were trained using the BCE loss function using gradient boosting backpropagation, and the conversion rate was predicted. Furthermore, each of the four corresponding functions had two different options (linear and nonlinear), so ensuring their adaptability was crucial.
[0079] Step 6: Model training and evaluation
[0080] The simplified feature matrix was fed into the LDNN and training continued, with the conversion rate as the target variable. By adjusting the training step size (0.01, 0.001, 0.0001) and the number of training cycles (500, 1000, 2000, 10,000), the model gradually learned the relationship between ligand structure, experimental conditions, and conversion rate.
[0081] To prevent overfitting, we used the dropout function, discarding some data at a time (6%, 12%, 24%, and 48%) to determine the optimal value for each training parameter. After training, we used the validation set to evaluate the model.
[0082] Assume that the conversion rate of the catalyst prepared by ligand A is predicted to be 70%, but the actual conversion rate is 50%, so the error is 10%. Through this evaluation, the model parameters are adjusted and determined as follows: (the corresponding function is Reul, the training step is 0.001, the number of training times is 2000, the dropout function parameter is set to 0.12, and the structure is as follows Figure 1 As shown in Figure 2), the probability that the final model’s prediction accuracy for different ligands was less than 5% reached more than 80%.
[0083] Step 7: Predict the conversion rate of unknown ligand
[0084] After training, the model can be used to predict the conversion of new ligands. Consider a new ligand X, whose composite SMILES string is C[P+]. RDKit generates a 256-bit fingerprint for ligand X, which is then combined with the corresponding experimental conditions into a feature matrix. For example, assuming the experimental conditions for X are a temperature of 210°C and an HCl ratio of 1.10, inputting its feature vector into the model yields a predicted conversion improvement of 70%. This output can help determine whether ligand X exhibits a good catalytic effect under the reaction conditions and warrants further experimental verification.
[0085] More than 8000 ligands were collected from https: / / www.chemsrc.com / and converted into SMLILES strings, with the central metal content set to 1%, the HCl ratio set to 1.15, and the acetylene space velocity set to 680 h -1 The reaction temperature was set to 180 degrees Celsius and the model was used for prediction. Approximately 100 ligands were obtained with excellent prediction results. After removing factors such as toxicity, danger, and price, two ligands were selected from the remaining ligands for verification.
[0086] Step 8: Prepare the catalyst
[0087] The ligand predicted to have good effect under the given reaction conditions in step 7 was used to prepare a catalyst with a central metal content of 1% Ru by impregnation method. The catalyst performance was tested experimentally using the same reaction conditions. At the same time, a catalyst with a central metal content of 1% Ru without adding ligand was prepared under the same conditions as a control to determine the improvement rate of the catalyst conversion rate after modification.
[0088] Step 9: Experimental Testing
[0089] For the predicted result C l = 0.71 ligand (CAS No.: 141556-45-8) and blank control were tested and the conclusion was drawn: r =88%, C e =61%, calculate C l =0.69, the effect is good.
[0090] For the predicted result C l = 0.80 ligand (CAS No.: 79917-90-1) and blank control were tested and the conclusion was drawn: r =94%, C e =61%, calculate C l =0.84, the effect is good.
[0091] The present invention is not limited to the specific technical solutions described in the above embodiments, and all technical solutions formed by equivalent replacement are within the protection scope required by the present invention.
Claims
1. A method for predicting the quality of acetylene hydrochlorination catalyst ligands based on deep neural networks, characterized in that: The method comprises the following steps: Step 1: Establish a data set and collect information on known catalyst ligands from the literature. The molecular structure of the ligand and the corresponding composite SMILES string, temperature, acetylene space velocity, central metal content, HCl ratio, and conversion rate are recorded in the database; Step 2: Define the catalyst conversion rate improvement Conversion rate improvement rate C l , specifically C e is the conversion rate of the catalyst control sample without ligand under the same conditions, C r is the conversion rate of the catalyst with added ligand; Step 3: Convert the molecular structure of the ligand into an ECFP fingerprint The composite SMILES string of each ligand is converted into a 256-bit binary vector called an "ECFP fingerprint." For known ligand SMILES strings, a search is performed to map each functional group to a molecular fingerprint consisting of 0s and 1s from the RDKit open-source library, combining them into a 256-bit molecular fingerprint. Different sequences of 0s and 1s represent different functional groups, thus reflecting the influence of different ligand structures. Step 4: Construct the feature tensor matrix Each ligand collected has its own 256-bit molecular fingerprint. The fingerprint data of all ligands are integrated into a tensor matrix. Each row of the matrix represents a ligand and each column is a feature bit. The experimental conditions temperature and acetylene space velocity are converted into vectors and added to the characteristic matrix; Again, the experimental conditions of HCl ratio and central metal content are converted into vectors and added to the feature matrix; Step 5: Build and tune the model architecture Using a deep neural network (DNN) algorithm, since the tensors containing airspeed and temperature differ significantly in magnitude, they are partitioned into separate input modules. A buffer layer is added as the first layer to reduce the differences between different variables. For each of the three input modules, the corresponding function Reul is used to convert them into three vectors. These are then concatenated into a matrix and converted into a 1*1 tensor using the Reul function again. The BCEloss function is used to train the parameters of the corresponding function using the gradient boosting method through backpropagation, and the conversion rate is predicted. Step 6: Model training and evaluation The simplified feature matrix group is input into the LDNN model and training continues. The target variable is the conversion rate. Through model training, the corresponding function is Reul, the training step size is adjusted to 0.001, and the number of training times is adjusted to 2000. The model gradually learns the relationship between ligand structure, experimental conditions and conversion rate. At the same time, in order to prevent overfitting, the dropout function is used to discard 12% of some data each time, and the optimal value of each training parameter is determined after trial and error; After training, the model was evaluated using the validation set. Finally, the model achieved a probability of over 80% that the difference between the predicted conversion rate and the actual conversion rate for different ligands was less than 5%. Step 7: Predict the conversion rate of unknown ligand After training, the model can be used to predict the conversion rate of a new ligand, ligand X. RDKit generates a 256-bit fingerprint for ligand X, which is then merged into a feature matrix along with the corresponding experimental conditions. After inputting the feature vector of X's experimental conditions, including temperature, HCl ratio, central metal content, and space velocity, the model outputs the predicted conversion improvement rate. This output can help determine whether ligand X has a good catalytic effect under the reaction conditions.
2. The method for predicting the quality of acetylene hydrochlorination catalyst ligands based on deep neural network according to claim 1 is characterized in that: The catalyst with a central metal content of 1% Ru was prepared by impregnation method, and the ligand was selected. As a modifier for acetylene hydrochlorination catalyst.