تجمعات النحاس والزنك ذات النواة المنخفضة على ZnZrOx تحفز تخليق الميثانول من CO2 Low-nuclearity CuZn ensembles on ZnZrOx catalyze methanol synthesis from CO2

المجلة: Nature Communications، المجلد: 15، العدد: 1
DOI: https://doi.org/10.1038/s41467-024-47447-6
PMID: https://pubmed.ncbi.nlm.nih.gov/38600146
تاريخ النشر: 2024-04-10

تجمعات النحاس والزنك ذات النواة المنخفضة على ZnZrOx تحفز تخليق الميثانول من CO2

تاريخ الاستلام: 23 سبتمبر 2023
تم القبول: 27 مارس 2024
نُشر على الإنترنت: 10 أبريل 2024
(ت) تحقق من التحديثات

ثايلان بينهيرو أراوجو جورجيوس جياناكاس © جوردي مورايس-فيدال (ب) ميخائيل أغراشيف زايرا رويز-برنال © فيل بريكشاس © تانغشينغ زو فرانك كرومايش © باتريك أ. ويلي، ويندلين ج. ستارك روبرت ن. غراس © غونار ييشكه © شانون ميتشل © نوريا لوبيز © وجافيير بيريز-راميريز

الملخص

يمكن أن يفتح تعزيز المعادن أداءً عاليًا في محفزات الزنك-الزركونيوم. ، من أجل الهدرجة إلى الميثانول. ومع ذلك، لا تزال الجهود المبذولة تركز على البالاديوم المكلف، ولا يزال اختيار المعدن الأمثل والهندسة المعمارية على المستوى الذري غير واضحين. هنا، نحقق في تعزيز محفزات بكميات صغيرة ( ) من معادن الهدرجة المتنوعة ( ، ، و Cu ) تم تحضيرها من خلال نهج موحد لرش اللهب البيروليز. يظهر النحاس كأكثر المحفزات فعالية، حيث يضاعف إنتاج الميثانول. تكشف تحليلات الامتصاص بالأشعة السينية أثناء التشغيل، والأشعة تحت الحمراء، وتحليل الرنين المغناطيسي الإلكتروني، ومحاكاة نظرية الكثافة الوظيفية أن تشكل الأنواع مجموعات CuZn منخفضة النواة وغنية بالزنك على السطح أثناء التفاعل، والذي يرتبط بتوليد فراغات الأكسجين في جوارها. تظهر الدراسات الآلية أن هذا التجمع الحفاز يعزز الهدرجة السريعة لل intermediates الفورمات إلى الميثانول بينما يثبط بشكل فعال إنتاج أول أكسيد الكربون، مما يبرز إمكانيات التجمعات المعدنية ذات النووية المنخفضة في تركيب الميثانول القائم على.

أكاسيد الزنك-الزركونيوم المختلطة ظهرت كعوامل حفازة فعالة من حيث التكلفة ووفيرة على الأرض لـ الهدرجة إلى الميثانول ، مادة كيميائية ذات صلة عالية وناقل للطاقة . هذه الأنظمة جذابة بشكل خاص نظرًا لقدرتها على قمع تكوين أول أكسيد الكربون غير المرغوب فيه (CO) من خلال تفاعل التحول العكسي للماء والغاز (RWGS) ( )، مما يؤدي إلى ارتفاع انتقائية الميثانول . بالإضافة إلى ذلك، تظهر المحفزات استقرارًا ملحوظًا أثناء التشغيل حتى عند تعرضها للشوائب مثل كبريتيد الهيدروجين وثاني أكسيد الكبريت، والتي توجد عادة في -تحتوي على مجاري . على الرغم من أن الترسيب المشترك (CP) تم تأسيسه كأول طريقة لتخليق محفزات أظهرت دراسة حديثة أن تقنية الرش اللهبي الحراري (FSP)
يحقق مواد ذات عائد زمني لمثانول أعلى بحوالي ثلاثة أضعاف مقارنة بالأنظمة المترسبة المشتركة . على الرغم من أن مواد CP و FSP تمتلك مجموعات تحفيزية مشابهة، إلا أن FSP تعظم المساحة السطحية وتكوين العزلات أنواع تقع في مواقع السطح من الشبكة، التي تعزز الأداء المحسن . في الواقع، تعزز الهندسة المعمارية الفريدة للمحفزات المصنوعة من اللهب إنشاء فراغات أكسجين على السطح، والتي تعد مكونات رئيسية للتجمعات النشطة التي تفضل تكوين الميثانول. ومع ذلك، بغض النظر عن طريقة التحضير، فإن التنشيط الهتروليتي لـ على هو الخطوة الأكثر تطلبًا ويحد من معدلات تكوين الميثانول .
ترقية المعادن هي استراتيجية راسخة لتعزيز قدرة انقسام الهيدروجين والأداء العام لمحفزات الأكسيد في
تحويل إلى الميثانول . بينما تم ترسيبها معًا من المعروف أن المحفزات تستفيد من إضافة كميات صغيرة من معادن الهدرجة، وتقتصر الأمثلة على عدد قليل من المعادن. ، والنحاس ) مع البالاديوم المعروف بأنه الأكثر فعالية . ومع ذلك، تطبيق هذه الاستراتيجية على اللهب المصنوع تعتبر الأنظمة التي تهدف إلى تحسين آفاق تنفيذها الصناعي تحديًا بسبب المعرفة المحدودة وقابلية نقل التأثيرات الترويجية من المحفزات التي تم تصنيعها تقليديًا. تعتمد تأثيرات معدن معين ليس فقط على هويته ولكن أيضًا على حالته الكيميائية، التي تتأثر بعوامل مختلفة، بما في ذلك طريقة التصنيع، ومحتوى المحفز، وإعادة بناء المحفز تحت ظروف التشغيل، وبنية الأكاسيد المستخدمة. على الرغم من التقدم الكبير في فهم تأثيرات الترويج على المحفزات الأخرى القائمة على الأكسيد، مثل أنظمة أكسيد الإنديوم دراسات متكاملة عن تفتقر المحفزات. لذلك، فإن استكشاف تعزيز اللهب المصنوع العوامل المساعدة لتوضيح هياكل التجمعات النشطة والأداء المرتبط بالأنظمة الناتجة، بما في ذلك تنوع المحفزات، وتأثيرها على تشكيل فراغات الأكسجين، والديناميات تحت ظروف التفاعل، أمر ضروري. على وجه الخصوص، تعتبر الدراسات التشغيلية أساسية لتأسيس علاقات قوية بين الهيكل والأداء لتوجيه تصميم التطبيقات العملية. محفزات .
في هذه الدراسة، نقوم بالتحقيق بشكل منهجي في تعزيز النار المصنوعة محفزات بواسطة المعادن الهيدروجينية ذات الصلة ، و Cu ) لـ الهدرجة إلى الميثانول. تكوين المحفزات القياسية وتقييمها بالإضافة إلى التوصيف المتعمق يكشف أن النحاس هو الأكثر فعالية كعامل محفز لـ ، حيث يؤدي إلى أكبر تحسين في الأداء. بالإضافة إلى ذلك، تم تعزيز النحاس تتكون المحفزات من بديل وفير في الأرض وفعال من حيث التكلفة للترويج التقليدي للبلاديوم، وهي متميزة تركيبيًا عن النحاس التقليدي. أنظمة، والتي تحتوي عادةً على من النحاس وتكون أقل انتقائية بشكل ملحوظ تجاه الميثانول تم تطبيق التحليلات المجهرية التفصيلية، والديناميكية، والاستقرار، وطيفية التشغيل، والمحاكاة النظرية لجمع الفهم حول تعزيز النحاس. على وجه التحديد، نحقق في الأنواع تحت ظروف التفاعل وتأثيرها على تشكيل فراغات الأكسجين، وهندسة الفرق النشطة، والتفاعل. توفر دراستنا رؤى قيمة حول الترويج بواسطة معادن مختلفة، ويقدم تبريرًا غير مسبوق على المستوى الذري لحالة العمل لـ محفزات لـ الهدرجة إلى الميثانول.

النتائج والمناقشة

أثر المحفزات على الأداء وهندسة المحفزات المعززة بالمعادن تم تصنيع المحفزات بواسطة تقنية الرش اللهبي الحراري (FSP، الجدول التكميلي 1)، والتي تم إثبات أنها تنتج مواد ذات أداء تحفيزي أفضل من النظائر المترسبة المشتركة لـ الهدرجة إلى
ميثانول . وبناءً عليه، فإن المحتوى الأمثل من الزنك ( تم اختيار ما تم الإبلاغ عنه لهذه الدراسة، في حين تم تخصيص مبلغ رمزي من تم اختيارها لجميع المروجين المعدنيين ( ، والنحاس)، لإنتاج محفزات بتركيبة مشابهة وبالتالي تمكين مقارنة مباشرة لتأثيراتها الترويجية على . كانت التركيبات التجريبية تتطابق بشكل وثيق مع القيم الاسمية، كما تم تحديده بواسطة مطيافية الانبعاث الضوئي البلازمي المقترن بالحث (ICP-OES، الجدول التكميلي 2). تقييم الأداء عند القيم ذات الصلة ظروف الهدرجة إلى الميثانول ، و ) أظهرت أن، بالنسبة لمعظم الأنظمة، يبقى عائد الميثانول الزمني المكاني (STY) تقريبًا دون تغيير (Re – و )، زادت قليلاً (حوالي ، و )، أو تحسن معتدل (حوالي ، رو- و ) بالمقارنة مع غير المروج ( ، الشكل 1). يمكن تتبع هذا الاتجاه في إنتاج الميثانول إلى التبادل المحدد بين تحويل ( ) وانتقائية الميثانول ( ) المعروضة بواسطة هذه المحفزات، مع Ru- و Pt-ZnZrO تحقيق أعلى (حوالي دون التعرض لفقدان كبير في (الشكل التوضيحي التكميلي 1). من المRemarkably، يُظهر النحاس أعلى تحسين في معدل إنتاج الميثانول (حوالي 120%، معدل الإنتاج = ، الشكل 1)، الذي يتجاوز ذلك الذي يسببه البالاديوم، المحفز الرائد الحالي لـ الإنتاجية اللافتة للميثانول مرتبط بقدرته على الحفاظ على أداء متميز (حوالي ) عند الارتفاع المستويات (حوالي الشكل التوضيحي التكميلي 1)، سلوك مشابه لوحظ أيضًا لـ . من المثير للاهتمام، غير المروج و العوامل المساعدة تظهر مساحة سطح محددة مشابهة جداً ) بشكل طازج (حوالي ، الجدول التكميلي 3 والشكل التكميلي 2) الذي يبقى تقريبًا دون تغيير بعد 20 ساعة من التوازن في التدفق. وهذا يشير إلى أن هذه الأنظمة مستقرة بشكل خاص تحت ظروف التفاعل وأن فرق أدائها لا ينشأ من اختلافات في . بالإضافة إلى ذلك، يبرز هذا مزايا استخدام FSP، حيث يسمح بالتخليق القابل للتكرار للعوامل الحفازة ذات التركيبات المتنوعة التي يمكن استخدامها ومقارنتها دون أي معالجة مسبقة (غسل أو تكليس) ودون الحاجة إلى تغيير إجراء التصنيع.
تم إجراء توصيف متعمق لتسليط الضوء على بنية المحفزات. أظهر التحليل بواسطة حيود الأشعة السينية (XRD، الشكل التكميلي 3) عدم وجود انعكاسات مميزة لمرحلة المعادن النقية أو أكسيد الزنك (ZnO) في المحفزات المستخدمة، مما يشير إلى أن المحفزات والزنك لا يزالان موزعين بشكل جيد أو يشكلان مراحل غير متبلورة. فيما يتعلق بـ ناقل، رباعي السطوح ( ) ” هو الطور البلوري السائد الموجود في جميع المواد الطازجة، والذي يتحول جزئيًا إلى الشكل أحادي الميل ( ) الهيكل عند التفاعل. من المحتمل أن يتم تحفيز هذا التحول بواسطة الماء المتكون تحت ظروف الهدرجة التحول الجزئي لـ إلى هو ظاهرة موثقة جيدًا للمواد المحتوية على الزركونيا التي تم إعدادها بواسطة . على النقيض، الأنظمة المبلغ عنها
الشكل 1 | الأداء المقارن للمعززات المعدنية محفزات.
عائد الميثانول الزمني، STY خلال الهدرجة ( من المعدن، ) المحفزات التي تم إعدادها بواسطة FSP. الخط الأفقي المتقطع يشير إلى
في الأدبيات، غالبًا ما يتم تحضيرها عن طريق الترسيب المشترك، مما يؤدي إلى دمج الزنك في الهيكل الكتلي لـ مكونًا محلولًا صلبًا يثبت الطور الرباعي، مما يمنع تحوله إلى الشكل أحادي الميل. بالمقارنة، عادةً ما تؤدي عملية FSP إلى مواد ذات مساحة سطح أعلى وتشتت زنك مُعظم في الطبقات الخارجية. . هذا يحسن من استخدام الزنك ويؤدي إلى استقرار أقل -الطور، الذي يمكّن من -إلى- تحول، يعزز من أداء الهدرجة.
أظهرت التحقيقات باستخدام مجهر الإلكترون الناقل الماسح (STEM) المرتبط بطيف الأشعة السينية المشتتة بالطاقة (EDX) أن جزيئات المعادن النانوية ذات الأحجام المتغيرة (التي تتراوح من 5 إلى 10 نانومتر) مرئية بوضوح للاستخدام. ، و المحفزات (الشكل 2 والشكل التكميلي 4). بالنسبة للذهب والبلاتين، تتوافق هذه الملاحظات مع تحليلات STEM-HAADF عالية الدقة، التي كشفت أيضًا أنه، على الرغم من عدم اكتشافها بواسطة خرائط EDX، فإن جزيئات نانوية بحجم حوالي 1 نانومتر موجودة في المحفزات التي تحتوي على الإيريديوم والبلاتين (الشكل التكميلي 5). بالمقابل، يبدو أن المعادن الأخرى مثل الرينيوم والكوبالت والنيكل والروثينيوم والنحاس موزعة بشكل جيد كأنواع ذات نووية منخفضة، كما تشير خرائط EDX، وفي حالة الرينيوم، أيضًا بواسطة STEM-HAADF (الشكل 2 والشكل التكميلي 4، 5). من المثير للاهتمام أنه بينما يبدو أن الزنك موزع ذريًا على جميع المحفزات، فإن جزءًا منه يميل إلى التجمع على في المناطق التي تحتوي على جزيئات نانوية من البالاديوم، مما يشير إلى أن هذه المراحل مرتبطة ومن المحتمل أن تشكل سبائك. الاختزال المبرمج بالحرارة مع الهيدروجين ( -TPR، الشكل التوضيحي الإضافي 6) يكشف أن معظم تظهر المحفزات بعض الإشارات المميزة ) دالة على تقليص المروجين، لا سيما لـ ، مما يشير إلى أن المعدن النقي أو يمكن أن تتشكل مراحل سبائك تحت ظروف التفاعل .
طيف هيكل امتصاص الأشعة السينية بالقرب من الحافة (XANES) لـ تظهر المحفزات المقاسة في وضع شبه في الموقع (عينات محمية في أنابيب تحت بيئة خالية من الهواء) اختلافات في نوعية المحفزات المعدنية (الشكل التوضيحي 7). الميزات الطيفية لـ Ag – و الأنظمة التي تم استكشافها في و الحواف، على التوالي، تتطابق بشكل وثيق مع تلك الخاصة بأفلام الفضة والذهب المعدنية، مما يشير إلى أن الفضة والذهب تشكل جزيئات نانوية معدنية كبيرة، بما يتماشى مع ملاحظات المجهر (الشكل التكميلي 4). بالمقابل، بينما تكون المحفزات في حالة معدنية في Pd – ، و العوامل المساعدة، من المحتمل أنها تتفاعل مع الزنك وربما تشكل سبائك، كما يشير إلى ذلك التحول في طيفها المقابل مقارنةً بمراجعها. ومع ذلك، فإن درجة السبائك مع الزنك منخفضة في عدة حالات، حيث يحتفظ الزنك في الغالب بطابعه الأكسيدي في جميعها. المحفزات (XANES، -الحافة، الشكل التوضيحي التكميلي 8). في الواقع، يمكن فقط تعيين مساهمات الزنك بشكل موثوق في طيف الامتصاص الدقيق للأشعة السينية الممتد (EXAFS، حافة الزنك K، الشكل التوضيحي 9) لـ . بالإضافة إلى ذلك، لا توجد إشارة مميزة لـ مسارات التشتت، عادة ما تكون مركزة عند حوالي. Åيمكن ملاحظته لأي ZnO العامل المساعد، مؤكدًا التشتت العالي لأنواع الزنك في معظم الأنظمة. أخيرًا، كانت المعادن مثل البلاتين، والرينيوم، والكوبالت، والنحاس في الغالب في حالة مؤكسدة (الشكل التوضيحي 7). من حيث المبدأ، نظرًا لأن المعادن كأيونات موجبة أقل عرضة للتفكك ، هذا يمكن أن يبرر
الأداء المحدود لبعض هذه العناصر كعوامل محفزة. ومع ذلك، فإنه من المثير للاهتمام بشكل خاص أن أنواع النحاس المؤكسد يمكن أن تقسم بشكل فعال ، الذي يتناقض مع قدرته المثبتة على تحسين إنتاجية الميثانول لـ (الشكل 1)، وبالتالي هناك حاجة لمزيد من التحقيق في تصنيفه تحت ظروف التفاعل.
بشكل عام، بالنظر إلى كل من النتائج المجهرية (الشكل 2، والشكل التكميلي 4، 5) والنتائج الطيفية (الشكل التكميلي 7-9)، لا توجد علاقة واضحة بين الأنواع وأداء يلاحظ وجود المحفزات، خاصة عند مقارنتها بـ الأنظمة التي تكون فيها الأنواع المعدنية الموزعة ذريًا أكثر فعالية كمحفزات من الكتل والجسيمات النانوية . بدلاً من ذلك، الترويج في يبدو أنه يعتمد أكثر على المعدن من اعتماده على الأنواع، حيث يظهر النحاس والبلاديوم تنوعًا مميزًا يؤدي إلى أداء مشابه وأفضل. في حالة البلاديوم، يكشف تحليل EXAFS لطيف حافة K للبلاديوم عن مساهمة واضحة من Pd-Zn (حوالي 6 جيران من الزنك، الشكل التوضيحي 10 والجدول التوضيحي 4)، مما يؤكد تشكيل جزيئات نانو سبيكة PdZn، ويتماشى مع TPR (الشكل التوضيحي الإضافي 6). نظرًا لأن البالاديوم هو محفز معروف لـ وتم توثيق سبائك PdZn جيدًا لخصائصها المواتية في تخليق الميثانول لقد خصصنا اهتمامًا خاصًا لجمع فهم أساسي لـ الترويج بواسطة النحاس، الذي يمثل فئة من المحفزات وفيرة الأرض، وأكثر استدامة، وغير مستكشفة.

تحديد نوع النحاس وتأثيره الترويجي

للحصول على رؤى أعمق حول أصل تعزيز النحاس في ، والمرجع المقابل و تمت دراسة المواد التي أعدتها FSP بشكل أعمق (الشكل 3 أ والشكل التكميلي 11، 12). تؤكد متانة المحفز التي تم تقييمها في اختبار مدته 100 ساعة (الشكل 3 أ) أن يظهر إنتاجية ميثانول أعلى بشكل ملحوظ، والتي تستقر بعد 50 ساعة من وقت التحفيز ( حوالي 0.55 ) مقارنةً بـ و المجمعة، مما يشير إلى أن أدائها لا ينتج عن تأثير إضافي بسيط من دمج كلا النظامين الثنائيين. فترة التحريض الملحوظة لـ العامل المساعد يتعلق بإعادة هيكلة وتوازن العامل المساعد عند درجات حرارة وضغوط مرتفعة. يتم ملاحظة سلوك مشابه لـ المحفز، على الرغم من أنه يتطلب وقتًا أقصر (حوالي 20 ساعة)، من المحتمل أن يكون ذلك بسبب البنية الأقل تعقيدًا لهذا المحفز. من المهم أن نلاحظ أنه بعد فترة الحضانة، يظهر أداءً مستقرًا، متجاوزًا بشكل كبير أداء الـ و الأنظمة. تم تحسين معدل إنتاج الميثانول مقارنة بـ و مرتبط بارتفاعه (حوالي 9 مقابل 4 و ، على التوالي) وفي حالة ، أيضًا (حوالي 90 مقابل )، كما يتضح من تقييم تم إجراؤه في ظروف مشابهة المستويات (حوالي 5٪، الشكل 3ب). من أجل الاكتمال، النحاس تمت أيضًا عملية التخليق وتم تقييم أدائها (الشكل التكميلي 11)، والذي يعادل لكنها أقل بكثير من ، مما يبرز الخصائص التحفيزية المتفوقة للنظام الثلاثي. ومن المثير للاهتمام أنه بينما تظهر جميع المحفزات ترتيب تفاعل ظاهري منخفض فيما يتعلق بـ لتركيب الميثانول (الشكل 3ج)، يعرض سلبية لتفاعل RWGS،
الشكل 2 | التنظيم الهيكلي لـ Pd و المحفزات. خرائط EDX لـ (أ) و (ب) العوامل المساعدة بعد الهدرجة لمدة 20 ساعة. ظروف التفاعل كما هو موضح في عنوان الشكل 1.
الشكل 3 | الاستقرار والحركية لـ المحفز. عائد الميثانول الزمني المكاني، STY، (ب) الانتقائية، خلال الهدرجة على ، و المحفزات. (ج) ترتيب التفاعل الظاهر بالنسبة لـ لـ الهدرجة إلى الميثانول (الأزرق) وتفاعل RWGS (البرتقالي). الخط الأفقي يشير إلى مجموع الميثانول من و محفزات.
تم تقييمه عند الثابت تحويل (حوالي ) عن طريق ضبط GHSV ( تم تقديم القيم المتوسطة المقاسة على مدى 20 ساعة في التدفق مع أشرطة الخطأ المقابلة لها. ظروف التفاعل كما هو موضح في عنوان الشكل 1.
مما يدل على أن تكوين أول أكسيد الكربون مثبط في ظروف الهدرجة المطبقة، مما يفسر انتقائيته العالية للميثانول. هذه النتائج تتماشى مع تجمع النحاس في جزيئات نانوية في عند التفاعل، والذي يُعرف بأنه يفضل تكوين CO وعلى الرغم من عدم اكتشافه بواسطة XRD (الشكل التكميلي 13)، إلا أنه يتضح بوضوح من خلال STEM-EDX (الشكل التكميلي 14a). في المقابل، يبقى النحاس موزعًا بشكل جيد كأنواع ذات نووية منخفضة في (الشكل التوضيحي 14ب)، حيث من المرجح أن تكون ذرات النحاس مرتبطة بالزنك، مما يشير إلى أن هذا النوع من التخصص يفضل تكوين الميثانول على إنتاج أول أكسيد الكربون.
لتسليط الضوء على تباين النحاس في تم إجراء تجارب امتصاص الأشعة السينية أثناء العمل (XAS) تحت المعالجة المسبقة في ومن 303 إلى 573 ك ظروف الهدرجة (1.5 ميغاباسكال و 573 كلفن) عند النحاس و -الحواف. تكشف تقنية XANES أثناء التشغيل أن كل من النحاس والزنك موجودان كأشكال مؤكسدة في العينة الجديدة العامل المساعد (الشكل 4 أ، ب). عند المعالجة الحرارية في الهيليوم، undergo كل من العنصرين تقليص جزئي، من المحتمل أن يشكلا و الأنواع. من المثير للاهتمام أن التحول إلى خليط التفاعل يسبب تغييرات مفاجئة في النحاس، ويظهر المزيد من الخصائص المعدنية، مما يتناقض مع النتائج شبه الحاضرة التي تظهر أنواع النحاس المؤكسد بعد التفاعل (الشكل 4 أ والشكل التكميلي 8). تسلط هذه التناقض الظاهر الضوء على أهمية الدراسات التشغيلية ويمكن أن تُعزى إلى الظواهر الأساسية مثل الطبيعة الشديدة الأكسوفيلية للنحاس، التي، في غياب بيئة مختزلة، يمكن أن تعزز إعادة أكسدته حتى بتركيزات صغيرة من الشوائب الأكسجينية التي توجد عادة في تيارات الهيليوم المستخدمة للحفاظ على العينات قبل التحليل. ومع ذلك، فإن الطيف
ميزات تحت ظروف التشغيل تختلف عن تلك الخاصة بـ ، الذي يحتوي على جزيئات نانوية معدنية من النحاس (الشكل 4أ). وهذا يشير إلى استقرار فعال للنحاس كأنواع معدنية ذات نووية منخفضة على المختلط السطح، على الأرجح بسبب تفاعل قوي مع الزنك. ومع ذلك، بينما يظهر الزنك مزيدًا من الاختزال عند تعرضه لمزيج التفاعل (الشكل 4ب)، إلا أنه لا يتم اختزاله بالكامل ولا يزال يحتفظ بخصائصه المميزة لـ ، مما يشير إلى أن نسبة صغيرة فقط من ذرات الزنك المعدنية تتشكل وتتفاعل مع النحاس، بينما قد تتكون الغالبية العظمى من أنواع الزنك المخفضة جزئيًا المختلطة على السطح. تظل هذه البنية سريعة التطور دون تغيير بعد 5 ساعات من التشغيل (الشكل 4 أ، ب)، مما يفسر الأداء المستقر لـ (الشكل 3أ).
تماشياً مع نتائج XANES، فإن تحليل EXAFS للعينة الجديدة و العوامل الحفازة تظهر تفاعلات قوية بين المعدن والأكسجين ) التفاعلات، مع حوالي 4 جيران من الأكسجين (الشكل 4c والجدول التكميلي 4). تحت الظروف المختزلة بشدة في بيئة الهدرجة، ومع ذلك، يتم ملاحظة فرق واضح بين هذين المحفزين. على وجه التحديد، المساهمات المشابهة لتلك الخاصة بورق النحاس موجودة على ، مع عدد كبير من الجيران النحاسيين (حوالي 9، الجدول التكميلي 4)، وهو ما يتماشى مع تشكيل أسطح معدنية ممتدة أو جزيئات نانوية. على النقيض من ذلك، بينما تظهر إشارات لـ ، هم بالتأكيد منتقلون من رقائق ولكن مع التمييز بين و التفاعلات مستحيلة بسبب عوامل التشتت المتشابهة (الشكل 4c). فيما يتعلق بحافة الزنك K، تظهر طيف EXAFS أن يعرض إلى حد كبير التفاعلات وفقط القليل يلاحظ أن المساهمة قد تم تحويلها مقارنةً بتلك الخاصة بورق الزنك (الشكل 4c). معًا مع XANES
الشكل 4 | تمييز النحاس والزنك في المحفز. أوبيراندو (أ) النحاس و (ب) -طيف XANES عند الحافة المحفز (1) خلال مرحلة التسخين معدل التسخين ، ووقت الإقامة 30 دقيقة)، (2) تحت الهيليوم عند 573 كلفن، وتحت ظروف التفاعل ( وقت الإقامة بعد (3) 30 دقيقة
و (4) 300 دقيقة على البث. (ج) طيف EXAFS المحول فوريه لـ العامل المساعد في شكله الطازج وتحت ظروف التفاعل التي تتوافق مع الأطياف في (أ) و (ب). أطياف XANES و EXAFS لأفلام النحاس والزنك، ZnO، والنحاس المنشط وقت الإقامة ) تظهر كمرجع.
تشير نتائج STEM-EDX (الأشكال 2 و 4a) إلى أن النحاس في حالة معدنية ومرتبط بقوة مع ذرات الزنك المخفضة جزئيًا، ومن المحتمل أن يشكل تجمعات CuZn غنية بالزنك ذات نووية منخفضة، خاصةً وأن محتوى الزنك الإجمالي أعلى بعشر مرات من محتوى النحاس، والعنصر الأخير موزع بشكل كبير.
تم إجراء محاكاة نظرية الكثافة الوظيفية (DFT) لتبرير الهياكل الأكثر احتمالاً الموجودة على ، و المحفزات. لهذا الغرض، تم فحص أكثر من 200 هيكل لتقييم درجة دمج النحاس والزنك في مواقع السطح، تشكيل فراغات الأكسجين، اتجاهات امتصاص النحاس، والتفاعل بين النحاس والزنك في المصفوفة. تم بناء النماذج استنادًا إلى معيارين: (1) التوافق مع الملاحظات التجريبية (طبيعة الكتلة)؛ و (2) كانت الطاقات المحسوبة باستخدام نظرية الكثافة الوظيفية معقولة (أي، العمليات المفضلة حراريًا). النماذج الأكثر صلة وطاقاتها المحتملة المقابلة (E) موضحة في الشكل 5، بينما يتم مناقشة النماذج الأخرى في المعلومات التكميلية (انظر الطرق والمناقشة التكميلية). يمثل بشكل أفضل نموذجًا يحتوي على ذرات زنك معزولة تحل محل مواقع الزركونيوم في سطح مع فراغ أكسجين قريب، ، الشكل 5 والشكل التكميلي 15) . بالمقابل، نموذجين مستقلين، أي و يتم النظر في المحفز، خاصةً منذ أن تظهر النتائج التجريبية وجود مرحلتين منفصلتين (الشكل التكميلي 14أ)، وكلاهما
دمج وامتصاص ذرة نحاس على تمتص الحرارة ( على التوالي، انظر الشكل 5). بالنسبة لـ ، استكشفنا في البداية دمج النحاس في نموذج مع مراكز زنك معزولة (الشكل 5). دمج النحاس ( ) غير مفضل ( فيما يتعلق بالنحاس الكمي ) حتى في المواقع القريبة من ذرات الزنك. ومن المثير للاهتمام، أن امتصاص النحاس على السطح بالقرب من مواقع الزنك ، الشكل 5) مفضل مقارنةً بـ خالي من الزنك ، مما يؤكد المزيد من التآلف. ومع ذلك، لم توفر أي من التكوينات التي تم التحقيق فيها مع ذرات الزنك المنفصلة مواقع نواة مناسبة للنحاس (الشكل 5). بالإضافة إلى ذلك، لا تتفق أي من النماذج تمامًا مع نتائج EXAFS و STEM-EDX، التي تشير إلى أن النحاس محاط على الأرجح بمجموعة غنية بالزنك تحت ظروف التفاعل (الجدول التكميلي 4). ومع ذلك، يجب تتبع تشكيل تجمعات CuZn تحت ظروف التفاعل إلى الألفة، نظرًا لطاقة التكوين للكتلة (النحاس الأصفر) وسبائك CuZn هي طاردة للحرارة و ، على التوالي). ومن ثم، تشمل السيناريوهات البديلة أن تشكيل الأزواج تحدث في وقت مبكر في التحضير. في الواقع، درجات حرارة الانصهار لـ و تختلف بشكل كبير (1023 و 2988 كلفن) مما يعني أن النواة أثناء تبريد عملية FSP يمكن أن تحدث بشكل مستقل. سيؤدي ذلك إلى بيئات غنية بالزنك مع النحاس ولاحقًا استقرار هذه الكتل بواسطة الأكسيد. مرحلة تضمن أن بعض ذرات الأكسجين ترتبط بذرات الزنك المحبة للأكسجين. وبالتالي، نحن
الشكل 5 | الوصف النظري لمواقع النحاس والزنك على العامل المساعد. لقطات لنماذج DFT مع طاقتها الكامنة النسبية المرتبطة بها، فيما يتعلق بالطاقة الكلية للنحاس والزنك و السطح، تم استخدامها لتبرير ميل الزنك والنحاس للاندماج في المواقع السطحية وترسيب النحاس على الزنك المخلوط والنقي الأسطح. النماذج المستخدمة لتقييم ، و تُبرز الأنظمة الحفازة من خلال محاكاة DFT باللون الأزرق الفاتح والرمادي والبنفسجي، على التوالي. الأسهم المتقطعة
تمثل دمج أو ذرات النحاس (باللون الأزرق والوردي الفاتح، على التوالي) مصحوبة بتكوين فراغات الأكسجين على السطح. تشير الخطوط الوردية الفاتحة الصلبة إلى ترسيب ذرة نحاس على و الأسطح. الشائع نمط موجود في و مُصوَّر باللون الأسود. رمز الألوان لنماذج DFT: Zn (أزرق)، Zr (أخضر)، Cu (وردي فاتح)، O (أحمر)، وعيوب الأكسجين (دوائر بيضاء منقطة).
نموذج تم التحقيق فيه حيث تتجمع 5 ذرات زنك في مواقع قريبة داخل سطح يحتوي على 5 فراغات أكسجين. استخدمنا هذه البنية لدراسة امتصاص ذرة نحاس واحدة ( ) العثور على أن هذا مفضل أكثر من و (الشكل 5). تضمين ذرة نحاس ثانية هو تفاعل طارد للحرارة ( ) وينتج عن ذلك نمط CuZn مشابه جدًا لذلك الموجود في النحاس الأصفر، لا سيما (110) السطح (الشكل 5). علاوة على ذلك، فإن تجمعات النحاس الأكبر (التي تحتوي على 3 أو 4 ذرات) ليست مفضلة (الشكل التكميلي 16) مما يبرز دور النواة وخصائص السبيكة في المركب النهائي. بالإضافة إلى ذلك، تم تقييم نماذج بديلة (أكثر من 150 هيكلًا) لاستكشاف متانة الـ مجموعة. قمنا بتقييم درجات مختلفة من الاختزال وتكوين هياكل مماثلة مع كميات مختلفة من الزنك والنحاس، كما هو موضح في المعلومات التكميلية (انظر الطرق والنقاش التكميلية، الشكل التكميلية 17-21، والجدول التكميلية 5). تشير هذه النتائج إلى أن النماذج التي تحتوي على 2 من النحاس (الذي قد يشكل أنواع ثنائية) و5 من الزنك هي الأكثر احتمالاً من الناحية الديناميكية الحرارية والأكثر توافقًا مع الملاحظات التجريبية. لذلك، تم اختيار الهيكل كنموذج الأكثر تمثيلاً لتقييم الخصائص الحفازة لـ العوامل المساعدة من منظور حسابي.

رؤى آلية حول تشكيل الفراغات الأكسجينية وديناميكياتها

لتحقيق في تشكيل وديناميات فراغات الأكسجين على النحاس تم إجراء تجارب مطيافية الرنين المغناطيسي الإلكتروني في الموقع (EPR) على هذا المحفز والمرجع و المواد (الشكل 6). طازج يظهر قمة ضعيفة تقريبًا ومتساوية الاتجاه ) وإشارة غير متساوية إضافية الشكل 6أ)، والتي تُنسب إلى الإلكترونات غير المتزاوجة المحصورة في فراغات الأكسجين المعزولة (تتصرف بشكل بارامغناطيسي، ) و المواقع، على التوالي، كما تم الإبلاغ عنه سابقًا عند المعالجة الحرارية في الهيليوم، زادت شدة كلا الإشارتين (الشكل 6أ)، مما يشير إلى تكوين كمية صغيرة نسبيًا من الفراغات الناتجة عن الحرارة. على النقيض من ذلك، فإن التعرض لـ
خليط التفاعل ( أدى إلى انخفاض في إشارة (الشكل 6 أ والشكل التوضيحي 22 أ). هذا يتماشى مع الملاحظات السابقة التي تشير إلى أن يتم إعادة أكسدته إلى عندما يتم نقل الإلكترونات إلى الفراغات ) تشكلت من تفاعل مع (انظر المناقشة التكميلية لوصف مفصل) . لا يزال، لا يؤثر بشكل فعال على خصائص مجموعة التحفيز، حيث يعمل كنوع متفرج. عند الانتقال إلى خليط التفاعل، تم ملاحظة إشارة جديدة عريضة ونُسبت إلى الإلكترونات غير المتزاوجة الموزعة على عدة فراغات أكسجين تتفاعل (بولارونات الفراغات المرتبطة بالتبادل). (الشكل 6 أ والشكل التوضيحي 22 أ). بشكل عام، يظهر هذا الإشارة الواسعة عندما تصبح كثافة الفجوات مرتفعة بشكل خاص، والتي تعتبر وصفًا رئيسيًا للأداء، كما تم الإبلاغ عنه لـ أنظمة . ل ومع ذلك، فقط و تكتشف الأنواع تحت المعالجة الحرارية في الهيليوم، بينما و لا تُلاحظ الإشارات (الشكل 6ب والشكل التوضيحي 22ب) حتى تحت ، مما يشير إلى أن تركيزًا منخفضًا من الشواغر موجود على الذي يمكن أن يفسر من حيث المبدأ أدائه الضعيف (الشكل 3أ).
طيف EPR للعينات الطازجة يعرض نموذجًا نموذجيًا إشارة، مع محور موتر وانقسام دقيق جيد التحديد لـ مكون (الشكل 6ج). من المثير للاهتمام، مثل ، لا تمت ملاحظة إشارة، تشير إلى أن المواقع هي أفضل لاقطات الإلكترونات. أدت المعالجة الحرارية في الهيليوم إلى انخفاض كبير في وزيادة في الإشارات، بينما ظهرت ميزة جديدة حول تداخل جزئي مع الـ “، ومتسق مع أيونات السوبر أوكسيد الممتصة على السطح (الشكل 6c). ومع ذلك، على عكس الإلكترونات التي يتم إطلاقها خلال هذه العمليات من المرجح أن تُحتجز بواسطة (i) المواقع، توليد EPR-silent الأنواع، (ii) الشواغر الفارغة، مما يؤدي إلى ، و (iii) الأكسجين المنطلق من الشبكة السطحية، مما يؤدي إلى تكوين السوبر أوكسيد. وهذا يدعم بشكل كبير أن وجوده يعزز بشكل ملحوظ تكوين الفراغات. التعرض لمزيج التفاعل يؤدي إلى الاختزال الكامل لـ وانخفاض كبير في إشارة (الشكل 6ج، د). بالإضافة إلى ذلك، إشارة مخصصة لـ تم اكتشافه ولكنه أوسع بكثير من الذي تم ملاحظته لـ ، مما يشير إلى أن كثافة أعلى من الشواغر هي
الشكل 6 | ديناميات فراغات الأكسجين خلال الهدرجة. طيف EPR في الموقع المقاس في درجة حرارة الغرفة لـ (أ) ، (ب) ، و (ج) العوامل المساعدة في شكلها الطازج، بعد تسخينها في الهيليوم معدل التسخين ) و الهدرجة ، ، و ). (د) طيف EPR أثناء التشغيل للإشارات الرئيسية لـ العامل المساعد تحت ،
، ووقت الإقامة ) وتحت ظروف التفاعل ( ، و ) مع الوقت المستمر على البث. تطور و أنواع، بالإضافة إلى فراغات الأكسجين البارامغناطيسية والفيرومغناطيسية و “على التوالي” لـ (هـ) ، (و) ، و (ز) المحفزات تحت ظروف التشغيل الموضحة في الشكل d والشكل التوضيحي 22a,b.
تم تشكيله في انخفاض الـ من المرجح أن يكون الإشارة مرتبطة بتحويلها إلى الميزة الواسعة ( )، حيث يؤدي التركيز المتزايد للفراغات إلى مواقع عيوب مرتبطة مغناطيسياً بشكل حديدي، وبالتالي، يستنفد عدد الفراغات المعزولة ( ).
لجمع رؤى إضافية حول ديناميات فجوات الأكسجين على تم إجراء قياسات EPR أثناء التشغيل تحت الهيليوم و ظروف الهدرجة (1 ميغاباسكال و
573 ك; الشكل 6د-ز، الشكل التكميلية 22أ، ب، و23). تطور الأنواع الرئيسية مع مرور الوقت (الشكل 6هـ-و) يكشف أن و تتطابق معدلات التكوين مع تلك الخاصة بـ و الخفض، مما يشير إلى أن هذه العمليات مرتبطة، وبالتالي يثبت آليات انتزاع الإلكترونات التي تم مناقشتها أعلاه. والأهم من ذلك، تظهر هذه التجارب أن -f الإشارات هي بصمات أفضل مقارنة بـ لمراقبة تشكيل الفراغات وإلغائها (انظر المناقشة التكميلية للحصول على التفاصيل). بالإضافة إلى ذلك، مقارنة لـ
ديناميكية اختزال و تظهر اختلافات ملحوظة (الشكل التوضيحي 24)، مع ملاحظة تدهور أحادي أسّي متكرر. . على النقيض، يوضح بوضوح مكونًا يت decay بنفس المعدل مثل في بينما يبدأ الآخر في الانخفاض في مرحلة لاحقة وبشكل أبطأ بكثير، وهو ما يرتبط على الأرجح بتقليل جزيئات النحاس النانوية الموجودة على ، بما يتماشى مع نتائج المجهر (الشكل التوضيحي 14أ). علاوة على ذلك، تشير هذه الملاحظات إلى أن الإجمالي التقليل أكثر كفاءة على يرجع ذلك على الأرجح إلى سلسلة من العوامل المترابطة مثل تنوعه والتفاعل القوي مع الشواغر من الزنك والأكسجين. ومن المثير للاهتمام، أن تعريض الـ المحفز للأكسجين في درجة حرارة الغرفة بعد التفاعل يكشف عن استعادة كاملة لـ إشارة (الشكل التوضيحي 25)، مما يشير إلى الطبيعة المستقرة لمواقع النحاس. بشكل عام، تُظهر تجارب EPR أن توليد فراغات الأكسجين يتزايد على عند التفاعل، والذي من المرجح أن يساهم في أدائه المحسن. في الواقع، تشير هذه النتائج بقوة إلى أن مواقع العيوب هذه هي مكون أساسي من مجموعات التحفيز الغنية بالزنك ذات النووية المنخفضة من CuZn الموجودة على ، حيث أن دورها متعدد الأوجه ومن المحتمل أن يشمل (ط) استقرار ذرات النحاس و/أو الوسائط التفاعلية، و(2) تنشيط المتفاعلات.

رؤى حول آلية التفاعل

يتطلب هدرجة الميثانول تفعيلًا فعالًا لـ و “، تليها نقل بروتون-هيدريد انتقائي لتجنب تكوين أول أكسيد الكربون من خلال تفاعل RWGS. آلية ثنائية الوظيفة تجمع بين خصائص الأكسيد ( التفعيل والهدرجة) والمعادن ( تكون مراحل (التفعيل) القريبة من بعضها البعض مفتاحًا لتجنب نقل الأنواع النشطة وتحقيق إنتاجية عالية من الميثانول. وبناءً عليه، يتم امتصاص المواد المتفاعلة الرئيسية، والوسائط، وأنواع المنتجات الثانوية ، و ) على النماذج المختارة المعروضة في الشكل 5 تم تقييمها من خلال محاكاة DFT لتبرير الخصائص الحفازة لـ ، و تفضل الامتزاز على المواقع القاعدية بينما التفاعل على هذه السطح ماص للحرارة (الشكل 7)، مع طاقة تنشيط ( ) من 0.70 إلكترون فولت، ويحدث بشكل غير متجانس، مكونًا هيدريدات وبروتونات ( ) على زوج (الشكل التوضيحي 26). على النقيض، يفعل بشكل سيء لكن هوموليتك يحدث الانفصال ) وامتصاص منتجاته (ذرات الهيدروجين) مفضل بشدة (الشكل 7). ومن ثم، فإن الانتقال من جزيئات النحاس النانوية إلى سطح، حيث يتم تنشيطه وهدرجته، مطلوب لـ العامل المساعد. وبالتالي، يؤدي هذا
إلى نقل بطيء للأنواع النشطة، مما يعيق تخليق الميثانول في هذا النظام، وبالتالي يساهم في أدائه الضعيف. من أجل heterوليتية تفعيل على الزوج مفضل أكثر و ) مقارنة بـ (الأشكال 7 و 9). بالإضافة إلى ذلك، يتم تنشيطه ككربونات في موقع قريب، مما يتجنب النقل غير الضروري للأنواع النشطة إلى سطح المحفز، مما يفسر إنتاجيته الأعلى من الميثانول مقارنةً بـ طاقة الامتزاز لـ و -المنتجات المنفصلة على نموذج، الذي يصف بشكل أفضل الفرق النشطة على النظام، يشير إلى أن هذا التكوين يعزز تنشيط هذه المتفاعلات في مواقع محددة مكانيًا، مما يتجاوز أيضًا النقل بعيد المدى للأنواع (الشكل 7). علاوة على ذلك، التفعيل خالٍ من الحواجز ويُظهر أنه يحدث بطريقة هوموليتية على ذرات النحاس المعدنية (الأشكال 7 و 9)، مما يدعم الدور الرئيسي للنحاس في تحسين الـ قدرة الانقسام . لوحظت نتائج مماثلة أيضًا للنماذج التناظرية التي تتضمن ذرات الزنك المخفضة جزئيًا المدمجة في مواقع الشبكة السطحية لـ ، فراغات الأكسجين، وذرات النحاس المعدنية المودعة بتوزيع مختلف ( -ب، الشكل التوضيحي الإضافي 27). على النقيض من ذلك، فإن طاقات الامتزاز للأنواع المتفاعلة على الـ تظهر السطح اتجاهات مشابهة لـ المفضل التفعيل وغير المفضل الامتزاز، مما يشير إلى أن الأنواع الزنك المخفضة جزئيًا، الموجودة على بدلاً من نظائرها المعدنية، هناك حاجة إلى تمكين تفعيل فعال والتحول اللاحق للمواد المتفاعلة إلى الميثانول (الشكل التوضيحي 27).
للحصول على رؤى تجريبية حول الأنواع الوسيطة التفاعلية الرئيسية التي تتشكل على ، و تم إجراء تجارب طيفية تحت ظروف التفاعل باستخدام تقنية الانعكاس المنتشر بالأشعة تحت الحمراء وتحويل فورييه (DRIFTS) (الشكل 8 والشكل التكميلي 28). عند التعرض لمزيج التفاعل ( “، و 573 كلفن؛ الشكل 8 أ-ج، والشكل التكميلي 28)، أظهرت جميع المحفزات نطاقات مميزة للفورمات ( ، “ ، و ) وأنواع الميثوكسي ( ، 1140، 2827 ، و )، والتي يمكن تخصيصها لوضعيات اهتزازية مميزة (الجدول التكميلي 6، 7) . هذه النتائج توضح أن تخليق الميثانول عبر الهدرجة على جميع المحفزات تتبع مسار الفورمات، وهو ما يتماشى مع الدراسات السابقة حول ويشير إلى ترويج النحاس في Cu لا تحكمه آلية مختلفة. ومن المثير للاهتمام، أن هناك إشارة قوية تُنسب إلى الطور الغازي (الشكل التوضيحي التكميلي 28) يتم اكتشافه لـ ، وهو ما يتماشى مع ميله لتعزيز هذا المنتج غير المرغوب فيه، مما يفسر انخفاضه
الشكل 7 | طاقات الامتزاز للأنواع التفاعلية. طاقات الامتزاز، لأنواع تفاعلية رئيسية على الأسطح ذات الصلة التي تمثل ، وCu أنظمة التحفيز. رمز اللون لنماذج DFT: Zn (أزرق)، Zr (أخضر)، Cu (فاتح
الوردي)، و O (الأحمر). بينما فقط الأكثر ملاءمة من الناحية الطاقية لأنواع التفاعلات الرئيسية لكل نموذج موضحة في الشكل 7، وهياكل الامتزاز الأخرى و وتم عرضها في الشكل التوضيحي 26 و27، على التوالي.
الشكل 8 | تطور الوسائط التفاعلية. طيف DRIFT في ظروف التشغيل للإشارات الرئيسية لـ (أ) ، (ب) ، و (ج) المحفزات تحت (أ-ج) الهدرجة ، ووقت الإقامة و (د-ف) بعد التحول إلى مزيج من متوازن مع
وقت الإقامة ). تطور تطور شدة الفورمات العادية ( ، برتقالي) و ميثوكسي (C ، أرجواني) لـ (ج) ، (ح) “، و (ط) المحفزات تحت ظروف التشغيل الموضحة في (أ-و).
مقارنة بـ و (الشكل 3ب). تطور و الأنواع التي تم تقييمها من خلال مراقبة نطاقات الاهتزاز عند 2874 و على التوالي، على مر الزمن-على البث (الشكل. ) كشفت أن تركيز سطحها يتغير في وقت واحد لجميع المحفزات، مما يشير إلى أن معدل التفاعل الكلي يتحدد بواسطة تركيز السطح. ومع ذلك، فإن تركيز السطح لـ و على يصل إلى حالة مستقرة في حوالي 30 دقيقة، بينما هو
يتطلب حوالي 50 إلى 70 دقيقة لـ و ، على التوالي (الشكل ). هذا يشير إلى أن المجموعة الحفازة في أكثر فعالية في تسريع تشكيل هذه الوسائط التفاعلية الرئيسية. للتحقيق في ديناميات الأنواع السطحية في النقاء تم استبداله بـ He وتم تطبيع شدة نطاقات الاهتزاز عند 2874 و تمت مراقبته باستمرار (الشكل 8d-i). تركيز ينخفض بسرعة، تدريجياً،
وببطء من أجل ، و ، على التوالي، على مدى 90 دقيقة على البث، مما يشير إلى حاجز طاقة أقل بكثير لعملية الهدرجة إلى على المحفز النحاسي محفز (شكل. ). ومع ذلك، على عكس و يتم هدرجة الشكل الجديد بسرعة نوع، يصل إلى حالة مستقرة مشابهة لتلك الموجودة في وجود كلاهما و (الشكل 8i). في هذه الحالة، فإن الهدرجة الإضافية لـ من المرجح أن يتم إخفاؤه بواسطة الامتصاص الانفصالي للميثانول على المواقع غير المشغولة التي أنشأها غياب .
للحصول على رؤى إضافية حول آلية التفاعل، قمنا بحساب ملفات الطاقة لـ الهدرجة إلى الميثانول وأول أكسيد الكربون على النماذج ذات الصلة المعروضة في الشكل 5 (الأشكال التكميلية 29، 30). اتبعنا مسار الفورمات، الذي يتماشى مع نتائج DRIFTS التشغيلية والدراسات السابقة لـ ومحفزات أخرى قائمة على أكاسيد المعادن . لقد قمنا أيضًا بتقييم تكوين الميثانول على السطح عبر هيدروجين CO حيث تم الكشف عن CO في الطور الغازي لـ و تكون عملية تكوين الميثانول مفضلة على إنتاج أول أكسيد الكربون في ، ، و (الشكل التوضيحي التكميلي 29-30)، بما يتماشى مع الملاحظات التجريبية (الشكل 3). على النقيض من ذلك، تم حساب ملفات الطاقة على تشير إلى الهدرجة إلى الميثانول تتنافس مع تكوين أول أكسيد الكربون. ومع ذلك، فإن إزالة أول أكسيد الكربون مفضلة على أن يتم هدرجة المزيد بما يتوافق مع المستوى الأدنى. من النحاس . علاوة على ذلك، تشير محاكاة DFT إلى الاستقرار النسبي بين و مفضل على فوق و (الجدول التكميلي 6). هذا يتماشى مع الهدرجة السريعة لـ إلى الميثانول الذي لوحظ على مقارنة بـ و ” (شكل. بشكل عام، استنادًا إلى DFT و DRIFTS، ننسب الإنتاجية العالية للميثانول التي تظهرها إلى تشكيل تجمعات CuZn منخفضة النواة غنية بالزنك مختلطة في المصفوفة. بشكل محدد، فإن هذا التجمع التحفيزي، الذي من المحتمل أن يتم استقراره بواسطة فراغات الأكسجين التي تتشكل عند التفاعل، يجمع بكفاءة بين كل من الأكسيدية ( ) وخصائص معدنية ( Cu )
مطلوب لتفعيل بشكل فعال و في المواقع القريبة وتمكين الهدرجة المتزامنة والسريعة لل intermediates الرئيسية إلى الميثانول (الشكل 9). أخيرًا، نظرًا للدور الرئيسي الذي تلعبه النحاس في تعزيز ، تحسين إضافي لمحتواه في النظام الثلاثي Cu يجب أن يتم تنفيذ الأنظمة في الأعمال المستقبلية.
باختصار، يتم الكشف عن النحاس كأكثر المحفزات فعالية لصنع اللهب. محفزات لتخليق الميثانول عبر الهدرجة، تنافس تعزيز البالاديوم المتطور، تتفوق على مختلف معادن الهدرجة (أي، ، ، أو Pt)، مما يؤدي إلى زيادة بمقدار الضعف في إنتاج الميثانول مقارنةً بالتحفيز غير المعزز . بالإضافة إلى ذلك، العامل المساعد يظهر أداءً أكثر استقرارًا وتفوقًا بشكل ملحوظ على مدى 100 ساعة من التشغيل مقارنةً بـ حيث يتجمع النحاس في جزيئات نانوية كبيرة، مما يؤدي إلى تفاعل أقل وانتقائية ميثانول ضعيفة. على النقيض من ذلك، يشكل النحاس أنواعًا معدنية ذات نووية منخفضة على عند التفاعل الذي يتفاعل بقوة مع مجموعات الزنك المختلطة في سطح الـ الناقل، كما يتضح من دمج تحليل XAS أثناء التشغيل ومحاكاة DFT. تظهر مطيافية EPR أثناء التشغيل أن توليد فراغات الأكسجين يتزايد بشكل خاص عند التفاعل ويرتبط ارتباطًا وثيقًا بتكوين تجمعات CuZn الغنية بالزنك. يتكامل التجمع الحفاز النشط الناتج الذي يتكون من تجمعات CuZn وفراغات الأكسجين في جوار ذرات الزنك بكفاءة مع الحمض والقاعدة و وظائف التقسيم التي تمكّن التفعيل والتفكك الهوموليتي بدون حواجز الانفصال على المواقع المكانية المحددة. في الواقع، هذه التكوين الهندسي الفريد يعزز بشكل كبير تكوين الميثانول من خلال الفورمات ( ) المسار، حيث يعزز الهدرجة السريعة لـ إلى الميثانول، مع منع النقل غير المرغوب فيه للأنواع النشطة وتكوين ثاني أكسيد الكربون. يحدد هذا العمل معلمًا مهمًا في تصميم المحفزات المعدنية. المحفزات وأنظمة أخرى تعتمد على أكاسيد قابلة للاختزال، كاشفة عن التأثير الترويجي لهندسة فعالة للغاية تعتمد على النحاس، وهو عنصر وفير في الأرض، من أجل الهدرجة إلى الميثانول. كما يبرز
الشكل 9 | مقارنة بين المرجع و المحفزات. ملخص للميزات الهندسية والتفاعلية الرئيسية التي تحدد أداء المحفزات. لقطات لامتصاص ومن ناتج التفكك
يتم عرض الامتزاز على أسطح مختلفة للأنظمة ذات الصلة. رمز الألوان لنماذج DFT: Zn (أزرق)، Zr (أخضر)، Cu (وردي فاتح)، O (أحمر)، H (أبيض)، وC (رمادي داكن).
إمكانات الكتل المعدنية المستقرة ذات النووية المنخفضة كفئة واعدة من المواد الحفازة غير المتجانسة لتطبيقات متنوعة ذات صلة.

طرق

تحضير المحفز

محفزات تحتوي على معزز معدني اسمي ومحتوى زنك بنسبة 0.5 و على التوالي، تم تحضيرها بواسطة FSP. باختصار، تم ضخ محلول سابق يحتوي على الزنك المذاب، والزركونيوم، ومعقدات المحفز (انظر الجدول التكميلي 1 للحصول على قائمة كاملة) بالنسبة المرغوبة من خلال فوهة بقطر 0.4 مم بمعدل تدفق قدره وت dispersed إلى رذاذ ناعم بواسطة تدفق الأكسجين عند ضغط 1.5 بار بمعدل تدفق . تم إشعال الرذاذ بواسطة لهب داعم تم توليده باستخدام الأكسجين و الميثان. لقد تم توصيف مثل هذه اللهب المولدة للجسيمات بشكل جيد وتم الإبلاغ عن وصولها إلى متوسط درجات حرارة اللهب ، مع معدلات تبريد سريعة جداً (حوالي تم جمع الجسيمات النانوية الناتجة على فلتر ألياف زجاجية (GF/A-6) واستخدامها في الهدرجة دون معالجة إضافية.

توصيف المحفز

تم إجراء مطيافية الانبعاث الضوئي للبلازما المقترنة بالحث (ICP-OES) باستخدام جهاز هوريبا ألترا 2 المزود بكاشف أنبوب الفوتومولتيبلاير. تم هضم العينات بمساعدة إشعاع الميكروويف باستخدام خليط من HCl (ألفا أيسار، ) ، (ألفا أيسر، )، و HF (سيغما ألدريتش، ) بنسبة حجم 2:1:0.5، تليها المعادلة باستخدام محلول مشبع من (فلوكا، ). تم إجراء امتصاص النيتروجين عند 77 كلفن باستخدام محلل Micromeritics TriStar II. قبل القياسات، تم إزالة الغازات من العينات عند 473 كلفن تحت فراغ لمدة 12 ساعة. المساحة السطحية المحددة ( تم تحديده باستخدام طريقة بروناور-إيميت-تيلر (BET). تم إجراء تحليل حيود الأشعة السينية (XRD) باستخدام جهاز حيود ريغاكوا سمارت لاب مع كاشف D/teX Ultra 250 باستخدام إشعاع Cu Ka. ) وتعمل في هندسة براج-برينتانو. تم الحصول على البيانات في نطاق بخطوة زاوية قدرها وزمن العد 1.5 ثانية لكل خطوة. الاختزال المبرمج حرارياً بالهيدروجين ( -TPR) تم قياسها عند الضغط الجوي باستخدام محلل Micromeritics AutoChem HP II. تم قياس تيار المخرج للجهاز باستخدام كاشف الموصلية الحرارية. تم إجراء تصوير الميكروسكوب الإلكتروني الناقل ذو المجال المظلم عالي الزاوية (HAADF-STEM) وتحليل الطيف بالأشعة السينية المشتتة للطاقة (EDX) على جهاز Talos F200X المزود بنظام Super-X EDS وتصوير STEM عالي الدقة على جهاز HD2700CS (Hitachi) المصحح للبروبة، وكلاهما يعمل عند 200 كيلوفولت. تم تفريق مسحوق المحفز في الإيثانول، وطحنه في هاون، وتفريقه على شبكة STEM من النحاس أو النيكل مغطاة بفيلم دعم من الكربون الشبكي. بعد تبخر الإيثانول في درجة حرارة الغرفة، تم تثبيت العينة على حامل مائل واحد وإدخالها في الميكروسكوب. تم أيضًا تحليل عينات مختارة بواسطة HAADF-STEM-EDX باستخدام Titan Themis المصحح للبروبة والذي يعمل عند 300 كيلوفولت. تم إجراء تجارب مطيافية امتصاص الأشعة السينية (XAS) في خطوط الأشعة السويسرية النرويجية (SNBL، BM31) في منشأة الإشعاع السنكروتروني الأوروبية (ESRF). تم تحديد شعاع الأشعة السينية باستخدام مقياس الطيف أحادي البلورة المزدوج المبرد بالنيتروجين السائل Si(111) تم قياس رقائق النحاس والزنك المعدنية من أجل المعايرة المطلقة للطاقة، باستخدام – غرف تأين مملوءة لمستويات امتصاص مثالية. تم استخدام تكوين هندسة الفلورية للتحليل تحت ظروف التشغيل، باستخدام كاشف Si SDD عنصر واحد مع تبريد بيلتير. تم إجراء مسح مستمر لكل من حواف Cu (بين 8880 و 9550 eV) و Zn (بين 9550 و 10650 eV)، وتم تعيين حجم الخطوة إلى 0.5 eV، مع مدة مسح قدرها 150 ثانية. تم تعيين حجم الشعاع إلى (أفقيًا عمودياً). المحفز ( ) تم وضعه بين سدادتين من صوف الكوارتز في خلية مفاعل شعرية من الكوارتز (قطر خارجي 1 مم، سمك جدار 0.01 مم). تم إجراء تجارب XAS في الخطوات التالية: (i) التسخين إلى 573 كلفن في معدل التسخين
، ووقت الإقامة )، (ii) التحويل إلى ظروف الهدرجة وقت الإقامة )، (iii) الانتقال إلى معدل التسخين ، ووقت الإقامة 30 دقيقة) و (iv) التبريد إلى درجة حرارة الغرفة في معدل التسخين ، ووقت الإقامة تم إجراء قياسات هيكل الامتصاص الدقيق للأشعة السينية الممتدة (EXAFS) في كل حالة مستقرة، بما في ذلك العينة الجديدة والمستخدمة، بالإضافة إلى تحت ظروف تفاعل الهدرجة. تحت جميع الظروف الأخرى، بما في ذلك تسخين درجة الحرارة (الخطوتان الأولى والرابعة) وبعد الانتقال إلى ظروف مختلفة (الخطوتان الثانية والرابعة، كما هو موضح أعلاه)، كانت قياسات هيكل الامتصاص القريب من الحافة بالأشعة السينية (XANES) مفضلة بدلاً من ذلك، لالتقاط التغيرات الديناميكية في الأنظمة. تم جرعات الغازات عالية النقاء بواسطة مجموعة من وحدات التحكم الرقمية في تدفق الكتلة من برونكهورست وتمت مراقبة النتيجة عبر الإنترنت من خلال مطياف الكتلة Pfeiffer Vacuum OmniStar GSD 320 O مع برنامج Quadera. تم معايرة الطيف الناتج للطاقة، وتصحيح الخلفية، وتطبيعها باستخدام برنامج Athena من مجموعة برامج Demeter. تم ملاءمة طيف EXAFS الموزون بشكل مثالي و -نوافذ (الجدول التكميلي 5) باستخدام برنامج أرتيميس. عامل تقليل السعة ( تم تحديد قيم 0.81 و 0.75 من خلال ملاءمة طيف EXAFS لرقائق النحاس والزنك، على التوالي. تم إنتاج مسارات التشتت للملاءمة باستخدام الهياكل البلورية المعروفة للنحاس المعدني، و ، و ZnO. تم إجراء قياسات XAS شبه في الموقع إضافية لـ Re و Ir و Pt حافة حافة، و ، و الحافة. تم إعداد العينات تحت استبعاد الهواء عن طريق نقل عينات المحفزات بعد الاختبارات الحفزية (20 ساعة تحت ظروف قياسية، انظر أدناه) إلى أنابيب زجاجية رقيقة من الكوارتز (قطر خارجي 1 مم، سمك جدار 0.01 مم، طول 80 مم، هيلغنبرغ GmbH) وإغلاق الأنابيب في جو خامل. تم وضع العينات على منصة بدلاً من الإعداد المستخدم لقياسات التشغيل وتم مسحها في وضع الفلورية باستخدام نفس كاشف Si SDD الموصوف أعلاه. تم إجراء تجارب مطيافية الرنين المغناطيسي الإلكتروني (EPR) في الموقع وفي التشغيل باستخدام إعداد مخصص (تردد الميكروويف الملعب المركزي عرض المسح سعة التعديل تردد التعديل طاقة الميكروويف تخفيف الطاقة وقت التحويل = 86.55 مللي ثانية ، الثابت الزمني ). عينة المحفز ( تم تحميله في شعيرة رملية من الكوارتز ) ووُضعت داخل أنبوب كوارتز EPR (ويلماد؛ تم وضع أنبوب EPR في مركز رنان عالي الحرارة مبرد بالماء مصنوع في المنزل. ، الذي تم تركيبه في مطياف EPR الموجي المستمر (CW) (Bruker EMX) الذي يعمل بترددات X-band. تم ضغط المفاعل إلى 1 ميغاباسكال، وتم تسخينه في تدفق الهيليوم ( ) إلى درجة الحرارة المطلوبة ( ) وتم الاحتفاظ بهذه الحالة لمدة ساعتين، مع تسجيل طيف EPR بشكل مستمر. ثم تم تبريده إلى درجة حرارة الغرفة وتم تسجيل طيف EPR. بعد ذلك، تم تسخين المفاعل مرة أخرى إلى 573 كلفن وترك لي stabilized لمدة 30 دقيقة. خليط التفاعل ( ثم تم قبوله وتم الحفاظ على تدفقه ) لمدة ساعتين. تم جرعة جميع الغازات بواسطة مجموعة من وحدات التحكم في تدفق الكتلة الرقمية وتم مراقبة مخرج المفاعل عبر الإنترنت باستخدام مطياف الكتلة Pfeiffer Vacuum Thermo-Star GSD 320 T1. تم الحصول على طيف EPR بشكل مستمر عند تدفق الغازات وتم تخزينه بشكل منفصل، باستخدام وضع اكتساب ثنائي الأبعاد، مما أتاح مراقبة زمنية دقيقة للعملية. تم تبريد المفاعل في النهاية وتم تسجيل طيف EPR في درجة حرارة الغرفة. ثم تم إدخال الهيليوم والهواء بعد ذلك وتم مراقبة طيف EPR حتى لم يتم الكشف عن أي تغييرات. تم إجراء مطيافية الانعكاس المنتشر بالأشعة تحت الحمراء بتحويل فورييه (DRIFTS) في إعداد مخصص يتكون من وحدة جرعة الغاز، ومنظمات الضغط الخلفي، ومطياف Pfeiffer Vacuum OmniStar GSD جهاز مطياف الكتلة لتحليل المنتجات عبر الإنترنت، وجهاز مطياف Bruker INVENIO-S مزود بكاشف من الزئبق والكادميوم والتيلوريوم (MCT) المبرد بالنيتروجين السائل. في تجربة نموذجية، تم تعبئة 50 ملغ من العينة غير المخففة في جهاز عالي-
غرفة تفاعل درجة الحرارة مع نوافذ ZnSe (هاريك ساينتيفيك)، والتي كانت مركبة في ملحق انعكاس منتشر على شكل صلاة مانتس (هاريك ساينتيفيك). بعد تطهير العينة بتدفق ، بان غاز، 4.6) لمدة 60 دقيقة، تم ضغط غرفة التفاعل إلى 1.5 ميغاباسكال وتم تسخينها إلى 573 كلفن مع زيادة تدريجية من . بعد ذلك، تم الحصول على الأطياف بشكل مستمر بدقة طيفية قدرها عن طريق تجميع 128 مسحًا كل 30 ثانية وتمت معالجتها مقابل خلفية KBr تم أخذها عند 573 كلفن و 1.5 ميغاباسكال. ثم تم تحويل تغذية الغاز إلى خليط من (بان غاز، 5.0) و في ، ميسر، 4.5) مع مولار نسبة 4 عند معدل تدفق إجمالي من . بعد 90 دقيقة تحت ظروف التفاعل، الـ توقف التدفق و متوازن مع تم إدخاله إلى غرفة التفاعل لمدة 90 دقيقة إضافية. تم معالجة البيانات الخام التي تم الحصول عليها باستخدام برنامج Bruker OPUS 8.2 وتم توليد طيف الاختلاف بالإضافة إلى ملفات كثافة باستخدام روتين بايثون مكتوب ذاتيًا، متاح كمصدر مفتوح على GitHub.https://github.com/philpreikschas/operando-ir).

تقييم المحفز

الهدرجة في الطور الغازي لـ تم إجراء التفاعل مع الميثانول باستخدام إعداد تدفق مستمر عالي الضغط من PID Eng&Tech يتكون من أربعة مفاعلات سرير ثابت متوازية، كما هو موضح في مكان آخر. محفزات غير مخففة (الكتلة، حجم الجسيمات تم تحميل ( ) في كل أنبوب تفاعل (قطر داخلي 4 مم)، مثبتة في مكانها بواسطة سرير من صوف الكوارتز موضوع على فريت الكوارتز، وتم تطهيرها في تدفق الهيليوم. بان غاز، 4.6) لمدة 30 دقيقة. تحت نفس التدفق، تم زيادة الضغط إلى 5 ميغاباسكال لاختبار التسرب، تلا ذلك تسخين سرير المحفز إلى تم إجراء التفاعل عن طريق تغذية مزيج من (بان غاز، 5.0 ) في ، ميسر، 4.5)، مع مولار نسبة 4 عند “، وسرعة الفضاء الغازية بالساعة (GHSV) من ما لم يُذكر خلاف ذلك. تم مقارنة انتقائية المحفزات عند درجة ثابتة من تحويل ( كما هو موضح في الشكل 3ب من خلال ضبط GHSV لكل نظام.
تم تحليل تيارات النفايات بواسطة الكروماتوغرافيا الغازية كل ساعة واحدة. عوامل الاستجابة ( ) لكل مركب ، بالنسبة للمعيار الداخلي ( فيه، ميسر، نقاء 3.5)، في تحليل الكروماتوغرافيا الغازية تم تحديده بواسطة المعادلة (1):
أين هي المنطقة المتكاملة المحددة لقمة المركب و هو تدفق المول المعروف المقابل عند مدخل المفاعل. تم استخدام متوسط 5 نقاط حول التركيز المتوقع للمحلول. تدفق المول غير المعروف لمركب تم تحديده باستخدام المعادلة (2):
أين هو التدفق المعروف لـ عند مخرج المفاعل.
تحويل ( ) ، الانتقائية ( ) ومعدل إنتاج الميثانول ( مеأن ) تم حسابها باستخدام المعادلات (3-5):
عائد الميثانول من حيث الزمن والمكان ( ) هو ناتج والوزن الجزيئي ( الميثانول ( ) كما هو مذكور في المعادلة (6):
تم تحديد توازن الكربون لكل تجربة وفقًا للمعادلة (7):
وكان دائمًا ضمن هامش.

التفاصيل الحاسوبية

تم إجراء محاكاة نظرية دالة الكثافة (DFT) باستخدام حزمة المحاكاة الأولية فيينا (VASP) ودالة الكثافة بيردو-بورك إرنزرهوف (PBE) تم توسيع الإلكترونات التكافؤية من مجموعة أساس الموجة المسطحة مع طاقة قطع حركية تبلغ 500 إلكترون فولت، بينما تم وصف الإلكترونات الأساسية باستخدام إمكانيات الزرع المعزز (PAW). . تم أخذ عينات من منطقة بريلوان باستخدام -شبكة مركزية بحجم شبكة متبادلة أضيق من Åالذي تم الحصول عليه بطريقة مونكهورست-باك معلمات الشبكة لـ ، و تم تحسينها بحد أدنى من قطع الطاقة الحركية يبلغ 700 إلكترون فولت. ثم، نماذج الشرائح لـ ، و تم فصل الإنهاءات من الكتل المحسّنة وتم تمثيلها بـ 2 ) ، ، ومع التوسعات، على التوالي. في جميع الحالات، تم تثبيت الطبقتين السفليتين وسمح للطبقتين الخارجيين بالاسترخاء. أضفنا منطقة فراغ من Åبين الألواح وتصحيح ثنائي القطب على طول تم تطبيق المحاور لأخذ عدم التماثل في الاسترخاءات بعين الاعتبار تم تقييم دمج الزنك والنحاس، بالإضافة إلى تكوين فراغات الأكسجين في مواقع مختلفة من وفقًا للإجراء المتبع في عملنا السابق . بالإضافة إلى ذلك، تم استكشاف ترسيب الذرات الفردية والعناقيد الصغيرة (حتى 4 ذرات) من النحاس في نماذج نقية ومخلوطة بالزنك. الأسطح. طاقات الامتزاز للجزيئات الرئيسية ( ، ، و تم تنفيذ ذلك على الأسطح المعنية لتبرير الخصائص الحفازة لـ ، و . تنظيف الأسطح والغازات ، وتم استخدام جزيئات CO كخزانات حرارية. تم امتصاص و تمت مع امتصاص ذرة هيدروجين في مواقع قريبة مثل OH أو لتجنب امتصاص الشظايا ذات الإلكترونات غير المتزاوجة التي تعدل الخصائص الإلكترونية لأكاسيد المعادن شبه الموصلة علاوة على ذلك، تم تقييم تأثير التفاعلات الضعيفة بعيدة المدى وخطأ التفاعل الذاتي على طاقات الامتزاز (الجداول التكميلية 8 و 9). تم تضمين تصحيحات التشتت من خلال نهج غريمي D3 بينما تم استخدام تصحيح هوبارد عن طريق نهج دوداريف مع تم تطبيق قيمة 4.00 إلكترون فولت على حالات ذرات الزركونيوم .
ملفات الطاقة لـ تم حساب الهدرجة إلى الميثانول والتفاعل التنافسي لتكوين أول أكسيد الكربون على نماذج مختارة من ، و العوامل المساعدة. الأسطح النظيفة وطور الغاز ، و تم استخدام الجزيئات كخزانات حرارية. صورة التسلق التي تدفع الشريط المرن (CI-NEB) ) تم استخدامه لتحديد حالات الانتقال. تم تأكيد طبيعتها من خلال حساب الترددات العددية بحجم خطوة Å.

توفر البيانات

البيانات المقدمة في الأشكال الرئيسية للمخطوطة متاحة للجمهور من خلال مستودع زينودو (الدخول: 8309938، الرابط: https:// doi.org/10.5281/zenodo.8309938) يمكن العثور على المدخلات والمخرجات لجميع محاكيات DFT عبر الإنترنت في مستودع ioChem-BD ، في https://iochem-bd.iciq.es/browse/review-collection/100/64621/fd52044500da8d298dc081fd. تتوفر بيانات إضافية تدعم نتائج هذه الدراسة في المعلومات التكميلية والمصدر.
البيانات. جميع بيانات المصدر الأخرى ذات الصلة متاحة من المؤلف المراسل عند الطلب. يتم تقديم بيانات المصدر مع هذه الورقة.

توفر الشيفرة

بيانات المصدر لتطبيق بايثون المستخدم في معالجة وتحليل بيانات الطيف الكهرومغناطيسي تحت التشغيل متاحة كمصدر مفتوح على GitHub (https://github.com/philpreikschas/operando-ir) ومتاحة أيضًا في مستودع زينودو (https://doi.org/10.5281/زينودو.10818472

References

  1. Li, K. & Chen, J. G. Hydrogenation to methanol over containing catalysts: insights into induced synergy. ACS Catal. 9, 7840-7861 (2019).
  2. Wang, J. et al. A highly selective and stable solid solution catalyst for hydrogenation to methanol. Sci. Adv. 3, 1-11 (2017).
  3. Han, Z. et al. hydrogenation to methanol on solid solution catalysts with ordered mesoporous structure. J. Catal. 396, 242-250 (2021).
  4. Pinheiro Araújo, T. et al. Design of flame-made catalysts for sustainable methanol synthesis from . Adv. Energy Mater. 13, 2204122 (2023).
  5. Pinheiro Araújo, T. et al. Methanol synthesis by hydrogenation of hybrid feeds. ChemSusChem 14, 2914-2923 (2021).
  6. Lee, K. et al. Atomic Pd-promoted solid solution catalyst for hydrogenation to methanol. Appl. Catal. B: Environ. 304, 120994 (2022).
  7. , Hong, X. & Liu, G. Highly dispersed metal doping to ZnZr oxide catalyst for hydrogenation to methanol: Insight into hydrogen spillover. J. Catal. 393, 207-214 (2021).
  8. Huang, C. et al. hydrogenation to methanol over PdZnZr solid solution: effects of the PdZn alloy and oxygen vacancy. ACS Appl. Energy Mater. 4, 9258-9266 (2021).
  9. Lee, K. et al. Engineering nanoscale H supply chain to accelerate methanol synthesis on . Nat. Commun. 14, 819 (2023).
  10. Pinheiro Araújo, T. et al. Flame-made ternary catalyst with enhanced oxygen vacancy generation for hydrogenation to methanol. Nat. Commun. 13, 5610 (2022).
  11. Pinheiro Araújo, T. et al. Flame spray pyrolysis as a synthesis platform to assess metal promotion in -catalyzed hydrogenation. Adv. Energy Mater. 12, 2103707 (2022).
  12. Wang, J. et al. Hydrogenation to methanol over -based catalysts: from mechanism to catalyst development. ACS Catal. 11, 1406-1423 (2021).
  13. Frei, M. S. et al. Atomic-scale engineering of indium oxide promotion by palladium for methanol production via hydrogenation. Nat. Commun. 10, 1-11 (2019).
  14. Bavykina, A. et al. Turning a methanation Co catalyst into an In-Co methanol producer. ACS Catal. 9, 6910-6918 (2019).
  15. Rui, N. et al. Hydrogenation of to methanol on a catalyst. ACS Catal. 10, 11307-11317 (2020).
  16. Snider, J. L. et al. Revealing the synergy between oxide and alloy phases on the performance of bimetallic In-Pd catalysts for hydrogenation to methanol. ACS Catal. 9, 3399-3412 (2019).
  17. Frei, M. S. et al. Nanostructure of nickel-promoted indium oxide catalysts drives selectivity in hydrogenation. Nat. Commun. 12, 1960 (2021).
  18. Kaiser, S. K. et al. Performance descriptors of nanostructured metal catalysts for acetylene hydrochlorination. Nat. Nanotechnol. 17, 606-612 (2022).
  19. Pinheiro Araújo, T. et al. Reaction-induced metal-metal oxide interactions in Pd- catalysts drive selective and Stable hydrogenation to methanol. Angew. Chem. Int. Ed. 62, 202306563 (2023).
  20. Beck, A. et al. Following the structure of copper-zinc-alumina across the pressure gap in carbon dioxide hydrogenation. Nat. Catal. 4, 488-497 (2021).
  21. Tsoukalou, A. et al. Operando X-ray absorption spectroscopy identifies a monoclinic :In solid solution as the active phase for the hydrogenation of to methanol. ACS Catal. 10, 10060-10067 (2020).
  22. Zabilskiy, M. et al. Mechanistic study of carbon dioxide hydrogenation over Pd/ZnO-based catalysts: the role of palladium-zinc alloy in selective methanol synthesis. Angew. Chem. Int. Ed. 60, 17053-17059 (2021).
  23. Prins, R. Hydrogen spillover: facts and fiction. Chem. Rev. 112, 2714-2738 (2012).
  24. Bahruji, H. et al. catalysts for direct hydrogenation to methanol. J. Catal. 343, 133-146 (2016).
  25. Zhao, H. et al. The role of species in single-atom catalyst for hydrogenation. Nat. Catal. 5, 818-831 (2022).
  26. Tada, S. et al. Ag addition to CuO-ZrO 2 catalysts promotes methanol synthesis via hydrogenation. J. Catal. 351, 107-118 (2017).
  27. Samson, K. et al. Influence of structure and copper electronic state on activity of catalysts in methanol synthesis from . ACS Catal. 4, 3730-3741 (2014).
  28. Ma, Y. et al. Reactivity of a zirconia-copper inverse catalyst for hydrogenation. J. Phys. Chem. C 124, 22158-22172 (2020).
  29. Wang, Y. et al. Strong evidence of the role of in affecting methanol selectivity from hydrogenation over . Chem 6, 419-430 (2020).
  30. Yang, C. et al. Strong electronic oxide-support interaction over for highly selective hydrogenation to methanol. J. Am. Chem. Soc. 142, 19523-19531 (2020).
  31. Wu, C. et al. Inverse as a highly efficient methanol synthesis catalyst from hydrogenation. Nat. Commun. 11, 5767 (2020).
  32. . et al. Insights into the interfacial structure of catalysts for methanol synthesis from hydrogenation: Effects of Cusupported nano-ZrO inverse interface. Chem. Eng. J. 470, 144006 (2023).
  33. Frei, M. S. et al. Mechanism and microkinetics of methanol synthesis via hydrogenation on indium oxide. J. Catal. 361, 313-321 (2018).
  34. Mädler, L. & Pratsinis, S. E. Bismuth oxide nanoparticles by flame spray pyrolysis. J. Am. Ceram. Soc. 85, 1713-1718 (2004).
  35. Kemmler, J. A. et al. Quenched, nanocrystalline high temperature phase for gas sensing applications. Sens. Actuators B: Chem. 161, 740-747 (2012).
  36. Meierhofer, F., Mädler, L. & Fritsching, U. Nanoparticle evolution in flame spray pyrolysis: process design via experimental and computational analysis. AIChE J. 66, 1-14 (2020).
  37. van Beek, W., Safonova, O. V., Wiker, G. & Emerich, H. SNBL, a dedicated beamline for combined in situ X-ray diffraction, X-ray absorption and Raman scattering experiments. Phase Transit. 84, 726-732 (2011).
  38. Ravel, B. & Newville, M. ATHENA, ARTEMIS, HEPHAESTUS: data analysis for X-ray absorption spectroscopy using IFEFFIT. J. Synchrotron Radiat. 12, 537-541 (2005).
  39. Zichittella, G., Polyhach, Y., Tschaggelar, R., Jeschke, G. & PérezRamírez, J. Quantification of redox sites during catalytic propane oxychlorination by operando EPR spectroscopy. Angew. Chem. Int. Ed. 60, 3596-3602 (2021).
  40. Pinheiro Araújo, T. et al. Impact of hybrid feeds on methanol synthesis over -based catalysts. Appl. Catal. B: Environ. 285, 119878 (2021).
  41. Kresse, G. & Furthmüller, J. Efficiency of ab-initio total energy calculations for metals and semiconductors using a plane-wave basis set. Comput. Mater. Sci. 6, 15-50 (1996).
  42. Kresse, G. & Furthmüller, J. Efficient iterative schemes for ab-initio total-energy calculations using a plane-wave basis set. Phys. Rev. B 54, 11169-11186 (1996).
  43. Perdew, J. P., Burke, K. & Ernzerhof, M. Generalized gradient approximation made simple. Phys. Rev. Lett. 77, 3865-3868 (1996).
  44. Blöchl, P. E. Projector augmented-wave method. Phys. Rev. B 50, 17953-17979 (1994).
  45. Kresse, G. & Joubert, D. From ultrasoft pseudopotentials to the projector augmented-wave method. Phys. Rev. B 59, 1758-1775 (1999).
  46. Monkhorst, H. J. & Pack, J. D. Special points for Brillouin-zone integrations. Phys. Rev. B 13, 5188-5192 (1976).
  47. Ricca, C., Ringuedé, A., Cassir, M., Adamo, C. & Labat, F. A comprehensive DFT investigation of bulk and low-index surfaces of polymorphs. J. Comput. Chem. 36, 9-21 (2015).
  48. Christensen, A. & Carter, E. A. First-principles study of the surfaces of zirconia. Phys. Rev. B 58, 8050-8064 (1998).
  49. Mizutani, U., Takeuchi, T. & Sato, H. Interpretation of the HumeRothery rule in complex electron compounds: -phase Alloy, FK-type and MI-type approximants. Prog. Mater. Sci. 49, 227-261 (2004).
  50. Jain, A. et al. Commentary: The materials project: a materials genome approach to accelerating materials innovation. APL Mater. 1, 011002 (2013).
  51. Makov, G. & Payne, M. C. Periodic boundary conditions in ab initio calculations. Phys. Rev. B 51, 4014-4022 (1995).
  52. Castelli, I. E. et al. Role of the band gap for the interaction energy of coadsorbed fragments. J. Phys. Chem. C 121, 18608-18614 (2017).
  53. Grimme, S., Ehrlich, S. & Goerigk, L. Effect of the damping function in dispersion corrected density functional theory. J. Comput Chem. 32, 1456-1465 (2011).
  54. Hubbard, J. Electron correlations in narrow energy bands II. the degenerate band case. Proc. R. Soc. A: Math. Phys. Eng. Sci. 277, 237-259 (1963).
  55. Dudarev, S. L., Botton, G. A., Savrasov, S. Y., Humphreys, C. J. & Sutton, A. P. Electron-energy-loss spectra and the structural stability of nickel oxide: An LSDA + U study. Phys. Rev. B 57, 1505-1509 (1998).
  56. Koga, H. et al. Facile NO-CO elimination over zirconia-coated 0) surfaces: Further evidence from DFT + U calculations. Appl Surf. Sci. 508, 145252 (2020).
  57. Syzgantseva, O. A., Calatayud, M. & Minot, C. Revealing the surface reactivity of zirconia by periodic DFT calculations. J. Phys. Chem. C 116, 6636-6644 (2012).
  58. Ruiz Puigdollers, A., Illas, F. & Pacchioni, G. Effect of nanostructuring on the reactivity of zirconia: a DFT + U study of au atom adsorption. J. Phys. Chem. C 120, 17604-17612 (2016).
  59. Tang, X., Zhang, H., Sun, C., Qiao, X. & Ju, D. Adsorption mechanisms over -modified surface for and CO): a DFT + U study. Surf. Sci. 716, 121976 (2022).
  60. Henkelman, G., Uberuaga, B. P. & Jónsson, H. A climbing image nudged elastic band method for finding saddle points and minimum energy paths. J. Chem. Phys. 113, 9901-9904 (2000).
  61. Pinheiro Araújo, T. et al. In situ formation of low-nuclearity CuZn ensembles on catalysts promotes hydrogenation to methanol [Data set]. Zenodo https://doi.org/10.5281/zenodo. 8309938 (2023).
  62. Álvarez-Moreno, M. et al. Managing the computational chemistry big data problem: the ioChem-BD platform. J. Chem. Inf. Model. 55, 95-103 (2015).
  63. Bo, C., Maseras, F. & López, N. The role of computational results databases in accelerating the discovery of catalysts. Nat. Catal. 1, 809-810 (2018).
  64. Preikschas, P. In situ formation of low-nuclearity CuZn ensembles on catalysts promotes hydrogenation to methanol [Code – philpreikschas/operando-ir: v0.1.1 (v0.1.1)]. Zenodo https:// doi.org/10.5281/zenodo. 10818472 (2024).

شكر وتقدير

تم إنشاء هذه المنشورة كجزء من NCCR Catalysis (رقم المنحة 180544)، وهو مركز وطني للكفاءة في البحث تم تمويله من قبل المؤسسة السويسرية الوطنية للعلوم. نشكر المركز العلمي لميكروسكوبية الضوء والإلكترون (ScopeM) في ETH زيورخ على الوصول إلى مرافقهم. نشكر السيد داريو فاوست أكل والسيد هنريك إلياسون على الحصول على بيانات HAADF-STEM-EDX. يشكر T.Z. وكالة العلوم والتكنولوجيا والبحث (A*STAR) في سنغافورة على الدعم من خلال زمالة دراسات عليا. يُعترف بوزارة العلوم والابتكار الإسبانية على الدعم المالي (PRE2019-088791، PID2021-122516OB-IOO، ومنحة سيفيرو أوتشوا MCIN/AEI/10.13039/501100011033 CEX2019-000925-S) ومركز برشلونة للحوسبة الفائقة – MareNostrum (BSC-RES) لتوفير موارد حاسوبية سخية. نشكر بول سانز بيرمان على دعمه في التصميم الجرافيكي. يُعترف بخطوط الأشعة السويسرية النرويجية (SNBL، ESRF) على توفير وقت الأشعة وموظفيها على الدعم القيم.

مساهمات المؤلفين

ت.ب.أ. و ج.ب.-ر. تصوروا ونسقوا جميع مراحل هذا البحث. ت.ب.أ.، ج.ج.، ج.م.ف، ن.ل.، س.م.، و ج.ب.-ر. كتبوا المقال بمشاركة جميع المؤلفين المشاركين الآخرين. ت.ب.أ. و ز.ر.ب. أجريا توصيف المحفزات وتقييمها. ج.م.ف و ن.ل. أجريا محاكاة نظرية الكثافة الوظيفية. ف.ك. أجرت تحليلات المجهر الإلكتروني. م.أ. و ج.ج. أجريا دراسات مطيافية الرنين المغناطيسي الإلكتروني. ج.ج. قيم بيانات مطيافية امتصاص الأشعة السينية. ب.أ.و.، ر.ن.ج.، و و.ج.س. أعدوا المحفزات باستخدام طريقة الرش اللهبي الحراري. ت.ز. أجرت تحليلات حركية. ب.ب. أشرف على الاستحواذ وقيم بيانات مطيافية الانعكاس المنتشر بالأشعة تحت الحمراء بتقنية تحويل فورييه. ج.ب.-ر. أشرف على المشروع بالكامل وأدار الموارد والتمويل. قدم جميع المؤلفين مدخلات للمخطوطة ووافقوا على النسخة النهائية.

المصالح المتنافسة

يعلن المؤلفون عدم وجود مصالح متنافسة.

معلومات إضافية

المعلومات التكميلية النسخة الإلكترونية تحتوي على
المادة التكميلية المتاحة على
https://doi.org/10.1038/s41467-024-47447-6.
يجب توجيه المراسلات وطلبات المواد إلى خافيير بيريز-راميريز.
معلومات مراجعة الأقران تشكر مجلة Nature Communications المراجعين المجهولين على مساهمتهم في مراجعة الأقران لهذا العمل. يتوفر ملف مراجعة الأقران.
معلومات إعادة الطبع والتصاريح متاحة علىhttp://www.nature.com/reprints
ملاحظة الناشر تظل Springer Nature محايدة فيما يتعلق بالمطالبات القضائية في الخرائط المنشورة والانتماءات المؤسسية.
الوصول المفتوح هذه المقالة مرخصة بموجب رخصة المشاع الإبداعي للاستخدام والمشاركة والتكيف والتوزيع وإعادة الإنتاج في أي وسيلة أو صيغة، طالما أنك تعطي الائتمان المناسب للمؤلفين الأصليين والمصدر، وتوفر رابطًا لرخصة المشاع الإبداعي، وتوضح ما إذا تم إجراء تغييرات. الصور أو المواد الأخرى من طرف ثالث في هذه المقالة مشمولة في رخصة المشاع الإبداعي للمقالة، ما لم يُذكر خلاف ذلك في سطر الائتمان للمادة. إذا لم تكن المادة مشمولة في رخصة المشاع الإبداعي للمقالة واستخدامك المقصود غير مسموح به بموجب اللوائح القانونية أو يتجاوز الاستخدام المسموح به، ستحتاج إلى الحصول على إذن مباشرة من صاحب حقوق الطبع والنشر. لعرض نسخة من هذه الرخصة، قم بزيارةhttp://creativecommons.org/licenses/by/4.0/.
© المؤلفون 2024

  1. معهد الكيمياء والهندسة الحيوية، قسم الكيمياء والعلوم الحيوية التطبيقية، ETH زيورخ، فلاديمير-بريلوغ-ويغ 1، 8093 زيورخ، سويسرا.
    معهد الأبحاث الكيميائية في كاتالونيا (ICIQ-CERCA)، معهد برشلونة للعلوم والتكنولوجيا، أفينيدا بايسوس كاتالان 16، 43007
    تاراغونا، إسبانيا. مختبر الكيمياء الفيزيائية، قسم الكيمياء والعلوم الحيوية التطبيقية، ETH زيورخ، فلاديمير-بريلوغ-ويغ 2، 8093 زيورخ، سويسرا. قسم الكيمياء غير العضوية ومعهد المواد (IUMA)، كلية العلوم، جامعة أليكانتي، ص.ب. 99، E-03080 أليكانتي، إسبانيا. مختبر الكيمياء غير العضوية، قسم الكيمياء والعلوم الحيوية التطبيقية، ETH زيورخ، فلاديمير-بريلوغ-ويغ 1، 8093 زيورخ، سويسرا. البريد الإلكتروني:jpr@chem.ethz.ch

Journal: Nature Communications, Volume: 15, Issue: 1
DOI: https://doi.org/10.1038/s41467-024-47447-6
PMID: https://pubmed.ncbi.nlm.nih.gov/38600146
Publication Date: 2024-04-10

Low-nuclearity CuZn ensembles on catalyze methanol synthesis from

Received: 23 September 2023
Accepted: 27 March 2024
Published online: 10 April 2024
(T) Check for updates

Thaylan Pinheiro Araújo , Georgios Giannakakis © , Jordi Morales-Vidal (B) , Mikhail Agrachev , Zaira Ruiz-Bernal © , Phil Preikschas © , Tangsheng Zou , Frank Krumeich © , Patrik O. Willi’, Wendelin J. Stark , Robert N. Grass © , Gunnar Jeschke © , Sharon Mitchell © , Núria López © & Javier Pérez-Ramírez

Abstract

Metal promotion could unlock high performance in zinc-zirconium catalysts, , for hydrogenation to methanol. Still, with most efforts devoted to costly palladium, the optimal metal choice and necessary atomic-level architecture remain unclear. Herein, we investigate the promotion of catalysts with small amounts ( ) of diverse hydrogenation metals ( , , and Cu ) prepared via a standardized flame spray pyrolysis approach. Cu emerges as the most effective promoter, doubling methanol productivity. Operando X-ray absorption, infrared, and electron paramagnetic resonance spectroscopic analyses and density functional theory simulations reveal that species form Zn -rich low-nuclearity CuZn clusters on the surface during reaction, which correlates with the generation of oxygen vacancies in their vicinity. Mechanistic studies demonstrate that this catalytic ensemble promotes the rapid hydrogenation of intermediate formate into methanol while effectively suppressing CO production, showcasing the potential of low-nuclearity metal ensembles in -based methanol synthesis.

Mixed zinc-zirconium oxides, , have emerged as cost-effective and earth-abundant catalysts for hydrogenation to methanol , a highly relevant platform chemical and energy vector . These systems are particularly attractive owing to their ability to suppress undesired carbon monoxide (CO) formation via the reverse water-gas shift (RWGS) reaction ( ), thereby leading to high methanol selectivity . Additionally, catalysts display remarkable on-stream stability even when exposed to impurities such as hydrogen sulfide and sulfur dioxide, which are commonly found in -containing streams . Although coprecipitation (CP) was established as the first method for synthesizing catalysts , a recent study revealed that flame spray pyrolysis (FSP)
yields materials with around threefold higher methanol space-time yield compared to coprecipitated systems . Despite CP and FSP materials possessing similar catalytic ensembles, FSP maximizes surface area and formation of isolated species located at surface positions of the lattice, which foster enhanced performance . Indeed, the unique architecture of the flame-made catalysts promotes the creation of surface oxygen vacancies, which are key components of the active ensembles that favor methanol formation. Regardless of the preparation method, however, the heterolytic activation of on is the most demanding step and limits methanol formation rates .
Metal promotion is a well-established strategy for enhancing the hydrogen splitting ability and overall performance of oxide catalysts in
the conversion of to methanol . While coprecipitated catalysts are known to benefit from the addition of small quantities of hydrogenation metals, examples are limited to a few metals ( , and Cu ), with palladium recognized as the most effective . However, applying this strategy to flame-made systems to improve their prospects for industrial implementation is challenging due to limited knowledge and transferability of promotional effects from conventionally synthesized catalysts. The effects of a given metal not only depend on its identity but also on its speciation, which is influenced by various factors, including the synthesis approach, promoter content, catalyst reconstruction under operating conditions, and structure of oxides used . Despite substantial progress in understanding promotional effects on other oxide catalysts, like indium oxide-based systems , integrated studies of catalysts are lacking. Therefore, the exploration of the promotion of flame-made catalysts to elucidate the active-ensemble structures and associated performance of the resulting systems, including promoter speciation, impact on oxygen vacancy formation, and dynamics under reaction conditions are essential. In particular, operando investigations are key for establishing robust structure-performance relationships to guide the design of practical catalysts .
In this study, we systematically investigate the promotion of flame-made catalysts by relevant hydrogenation metals ( , and Cu ) for hydrogenation to methanol. Standardized catalyst synthesis and evaluation as well as in-depth characterization reveal copper as the most effective promoter for , as it leads to the largest performance improvement. In addition, Cu -promoted catalysts comprise an earth abundant and cost-effective alternative to conventional palladium promotion and are compositionally distinct from traditional Cu – systems, which generally contain of Cu and are significantly less selective to methanol . Detailed microscopy, kinetic, stability, operando spectroscopy analyses, and theoretical simulations are applied to gather understanding of copper promotion. Specifically, we investigate the speciation under reaction conditions and its effect on oxygen vacancy formation, the architecture of active ensembles, and reactivity. Our study provides valuable insights into promotion by different metals, and offers unprecedented atomic-level rationalization of the working state of catalysts for hydrogenation to methanol.

Results and discussion

Impact of promoters on performance and catalyst architecture Metal-promoted catalysts were synthesized by flame spray pyrolysis (FSP, Supplementary Table 1), which was previously demonstrated to produce materials with better catalytic performance than coprecipitated analogues for hydrogenation to
methanol . Accordingly, the optimal zinc content ( ) reported was selected for this study, whereas a nominal amount of was chosen for all metal promoters ( , and Cu ), to produce catalysts with similar composition and thus enable a direct comparison of their promotional effects on . Both experimental compositions closely matched nominal values, as determined by inductively coupled plasma optical emission spectroscopy (ICP-OES, Supplementary Table 2). Performance evaluation at relevant hydrogenation to methanol conditions ( , and ) showed that, for most systems, methanol space-time yield (STY) remains virtually unchanged ( Re – and ), slightly increased (ca. , and ), or moderately improved (ca. , Ru- and ) in comparison to unpromoted ( , Fig. 1). This trend in methanol productivity can be traced back to the specific trade-off between conversion ( ) and methanol selectivity ( ) displayed by these catalysts, with Ru- and Pt-ZnZrO attaining higher (ca. ) without experiencing significant loss in (Supplementary Fig. 1). Remarkably, Cu exerts the highest improvement in the methanol STY (ca. 120%, STY= , Fig. 1), which surpasses that induced by Pd, the current state-of-the-art promoter for . The striking methanol productivity of is linked to its ability to sustain an outstanding (ca. ) at high levels (ca. Supplementary Fig. 1), a similar behavior also observed for . Interestingly, unpromoted and catalysts show very similar specific surface area ( ) in fresh form (ca. , Supplementary Table 3 and Supplementary Fig. 2) which remains virtually unaltered after 20 h equilibration on stream. This indicates that these systems are particularly stable under reaction conditions and their performance difference does not stem from differences in . In addition, this highlights the advantages of using FSP, as it allows the reproducible synthesis of catalysts of varying compositions that can be used and compared without any pre-treatment (washing or calcination) and having to change the manufacturing procedure.
In-depth characterization was conducted to shed light on the architecture of catalysts. Analysis by X-ray diffraction (XRD, Supplementary Fig. 3) revealed no reflections characteristic of pure metal phases or zinc oxide ( ZnO ) in the used catalysts, indicating that the promoters and zinc remain well dispersed or form amorphous phases. Concerning the carrier, tetragonal ( ) is the predominant crystalline phase present in all fresh materials, which partially transforms into the monoclinic ( ) structure upon reaction. This transformation is likely triggered by water formed under hydrogenation conditions . The partial transformation of into is a well-documented phenomenon for zirconia-containing materials prepared by . In contrast, systems reported
Fig. 1 | Comparative performance of metal-promoted catalysts.
Methanol space-time yield, STY during hydrogenation ( of metal, ) catalysts prepared by FSP. The horizontal dashed line indicates the
in the literature are often prepared by coprecipitation, which leads to the incorporation of zinc into the bulk structure of , forming a solid solution that stabilizes the tetragonal phase, preventing its transformation into the monoclinic form. Comparatively, FSP typically results in materials with higher surface area and maximized zinc dispersion in the outmost layers of . This improved Zn utilization and leads to a less stabilized -phase, which enables the thermodynamically favored -to- transformation, enhancing its hydrogenation performance.
Investigations by scanning transmission electron microscopy (STEM) coupled to energy-dispersive X-ray (EDX) spectroscopy showed that metal nanoparticles with variable sizes (ranging from 5 10 nm ) are clearly visible for used , and catalysts (Fig. 2 and Supplementary Fig. 4). For Au and Pd , these observations are corroborated by high-resolution STEM-HAADF analyses, which further revealed that, despite being undetected by EDX maps, nanoparticles of ca. 1 nm are present in Ir – and Pt -containing catalysts (Supplementary Fig. 5). In contrast, other metals like Re, Co, Ni, Rh, and Cu seem well dispersed as low-nuclearity species, as suggested by EDX maps, and, in the case of Re, also by STEM-HAADF (Fig. 2 and Supplementary Fig. 4,5). Interestingly, while zinc appears atomically dispersed over all catalysts, part of it tends to aggregate on at regions containing palladium nanoparticles, suggesting that these phases are associated and likely form alloys. Temperatureprogrammed reduction with hydrogen ( -TPR, Supplementary Fig. 6) reveals that most catalysts exhibit some distinguishable signals ( ) indicative of the reduction of promoters, particularly for , suggesting that pure metallic or containing alloys phases can be formed under reaction conditions .
X-ray absorption near-edge structure spectra (XANES) of catalysts measured in quasi in situ mode (samples protected in capillaries under air-free environment) evidence differences in the metal promoters’ speciation (Supplementary Fig. 7). Spectral features of the Ag – and systems probed at the and edges, respectively, closely match those of metallic Ag and Au foils, indicating that Ag and Au form large metallic nanoparticles, in line with microscopy observations (Supplementary Fig. 4). In contrast, while promoters are in a metallic state in Pd -, , and catalysts, they are likely interacting with zinc and possibly forming alloys, as suggested by the shift in their corresponding spectra compared to that of references. Still, the degree of alloying with zinc is low in several cases, as zinc mostly retains its oxidic character in all catalysts (XANES, -edge, Supplementary Fig. 8). Indeed, Zn contributions can only be reliably assigned in the extended X-ray absorption fine structure spectra (EXAFS, Zn K-edge, Supplementary Fig. 9) of . Additionally, no signal characteristic of scattering paths, generally centered at ca. for ZnO , could be observed for any catalyst, confirming the high dispersion of zinc species in most systems. Finally, the metals such as Pt, Re, Co, and Cu were mostly in oxidized state (Supplementary Fig. 7). In principle, since metals as cations are less prone to split , this could rationalize
the limited performance of some of these elements as promoters. Still, it is particularly intriguing that oxidic copper species can effectively split , which contrasts with its proven ability to improve the methanol productivity of (Fig. 1), and therefore further investigation into its speciation under reaction conditions is required.
Overall, considering both microscopic (Fig. 2, and Supplementary Fig. 4,5) and spectroscopic (Supplementary Fig. 7-9) findings, no clear correlation between the speciation and the performance of catalysts is observed, particularly when compared to systems in which atomically-dispersed metal species are more effective promoters than clusters and nanoparticles . Instead, promotion in appears more metal-dependent than speciation-dependent, with copper and palladium displaying distinct speciation that lead to similar and superior performance. In the case of Pd, EXAFS fitting of the Pd K-edge spectra reveals a clear Pd-Zn contribution (ca. 6 Zn neighbors, Supplementary Fig. 10 and Supplementary Table 4), confirming the formation of PdZn nanoalloy particles, and in line with TPR (Supplementary Fig. 6). Since palladium is a well-known promoter for and PdZn alloys are well documented for their favorable methanol synthesis properties , we have devoted special focus to gather fundamental understanding of promotion by copper, which represents an earth-abundant, more sustainable, and unexplored promoter class.

Copper speciation and promotional effect

To gain deeper insights into the origin of copper promotion in , and corresponding reference and materials prepared by FSP were further investigated (Fig. 3a and Supplementary Fig. 11,12). Catalyst durability assessed in a 100 -h test (Fig. 3a) confirms that shows a significantly higher methanol productivity, which stabilizes after 50 h induction time ( ca. 0.55 ) compared to that of and combined, indicating that its performance does not result from a simple additive effect of integrating both binary systems. The induction period observed for the catalyst relates to the restructuring and equilibration of the catalyst at elevated temperature and pressure. A similar behavior is observed for the catalyst, albeit requiring a shorter time (ca. 20 h ), likely due to the less complex architecture of this catalyst. It is important to note that beyond the induction period, demonstrates stable performance, greatly surpassing that of the and systems. The improved methanol STY of compared to and is linked to its higher (ca. 9 versus 4 and , respectively) and, in the case of , also (ca. 90 versus ), as evidenced by evaluation carried out at similar levels (ca. 5%, Fig. 3b). For completeness, Cu was also synthesized and its performance evaluated (Supplementary Fig. 11), which is comparable to but notably inferior to , further highlighting the superior catalytic properties of the ternary system. Interestingly, while all catalysts exhibit a low apparent reaction order with respect to for methanol synthesis (Fig. 3c), displays a negative for the RWGS reaction,
Fig. 2 | Structural organization of Pd- and catalysts. EDX maps of (a) and (b) catalysts after hydrogenation for 20 h . Reaction conditions as in the caption of Fig. 1.
Fig. 3 | Stability and kinetics of the catalyst. a Methanol space-time yield, STY, (b) selectivity, during hydrogenation over , and catalysts. (c) Apparent reaction order with respect to for hydrogenation to methanol (blue) and the RWGS reaction (orange). The horizontal line indicates the sum of the methanol of and catalysts.
was assessed at constant conversion (ca. ) by adjusting the GHSV ( ). Averaged values measured over 20 h on stream are presented with their corresponding error bars. Reaction conditions as in the caption of Fig. 1.
indicating that CO formation is inhibited over at the applied hydrogenation conditions, rationalizing its high methanol selectivity. These findings are in line with copper agglomerating into nanoparticles in upon reaction, which is known to favor CO formation and, although not detected by XRD (Supplementary Fig. 13), it is clearly evidenced by STEM-EDX (Supplementary Fig. 14a). In contrast, copper remains well dispersed as low-nuclearity species in (Supplementary Fig. 14b), as copper atoms are most likely associated with zinc, indicating that this speciation favors methanol formation over CO production.
To shed light into the copper speciation in the catalyst at work, operando X-ray absorption spectroscopy (XAS) experiments were conducted under pretreatment in and from 303 to 573 K ) and hydrogenation conditions ( 1.5 MPa and 573 K ) at the Cu and -edges. Operando XANES reveals that both copper and zinc exist as oxidized phases in the fresh catalyst (Fig. 4a, b). Upon thermal treatment in He , both elements undergo partial reduction, likely forming and species. Interestingly, switching to the reaction mixture causes abrupt changes to copper, and more metallic character is evidenced, which contrasts the quasi in situ findings that show oxidized copper species after reaction (Fig. 4a and Supplementary Fig. 8). This apparent discrepancy highlights the importance of operando studies and can be attributed to underlying phenomena such as the highly oxophilic character of copper, which, in the absence of a reducing environment, can promote its reoxidation even by small concentration of oxygen impurities typically present in He streams used for preserving samples before analysis. Still, spectral
features of under operating conditions deviate from those of , which contains metallic Cu nanoparticles (Fig. 4a). This hints at an effective stabilization of copper as low-nuclearity metallic species on the mixed surface, owing most likely to a strong interaction with zinc. Nonetheless, while zinc shows further reduction when exposed to the reaction mixture (Fig. 4b), it is not fully reduced and still maintains characteristic features of , suggesting that only a small fraction of metallic zinc atoms forms and interacts with copper, while its majority may comprise partially reduced zinc species intermixed on the surface. This rapidly evolving architecture remains unchanged after 5 h on stream (Fig. 4a, b), accounting for the stable performance of (Fig. 3a).
In line with XANES findings, EXAFS analysis of the fresh and catalysts exhibits strong metal-oxygen ( ) interactions, with ca. 4 oxygen neighbors (Fig. 4c and Supplementary Table 4). Under the highly reducing hydrogenation environment, however, a clear difference is observed between these two catalysts. Specifically, contributions similar to those of Cu foil are present on , with a large number of copper neighbors (ca. 9, Supplementary Table 4), which is consistent with the formation of metallic extended surfaces or nanoparticles. In contrast, while signals are evidenced for , they are clearly shifted from foil but distinguishing between and interactions is impossible due to their similar scattering factors (Fig. 4c). Regarding the Zn Kedge, the EXAFS spectra shows that largely exhibits interactions and only a small contribution is observed, which is shifted compared to that of the Zn foil (Fig. 4c). Together with XANES
Fig. 4 | Copper and zinc speciation in the catalyst. Operando (a) Cu and (b) -edge XANES spectra of catalyst (1) during the heating ramp in , heating rate , and dwell time 30 min ), (2) under He at 573 K , and under reaction conditions ( , dwell time ) after (3) 30 min
and (4) 300 min on stream. (c) Fourier-transformed EXAFS spectra of catalyst in fresh form and under reaction conditions corresponding to the spectra in (a) and (b). XANES and EXAFS spectra of Cu and Zn foils, ZnO , and activated Cu , dwell time ) are shown as reference.
and STEM-EDX results (Figs. 2 and 4a), these observations indicate that copper is in a metallic state and strongly associated with partially reduced zinc atoms, most likely forming low-nuclearity zinc-rich CuZn clusters, particularly since the overall zinc content is 10 times higher than that of copper and the latter element is highly dispersed.
Density functional theory (DFT) simulations were performed to rationalize the most likely architectures present on , and catalysts. For this purpose, more than 200 structures were examined to evaluate the degree of incorporation of Cu and Zn into surface sites, formation of oxygen vacancies, Cu adsorption trends, and interaction between Cu and Zn in the matrix. Models were built based on two criteria: (i) the compatibility with the experimental observations (cluster nature); and (ii) the DFT-computed energies were reasonable (i.e., processes thermodynamically favored). The most relevant models and their corresponding potential energies (E) are depicted in Fig. 5, whereas other models are discussed in the Supplementary Information (see Supplementary Methods and Discussion). is best represented by a model with isolated Zn atoms replacing Zr sites at the surface with a nearby oxygen vacancy, , Fig. 5 and Supplementary Fig. 15) . In contrast, two independent models, i.e. and , are considered for the catalyst, particularly since experimental results evidence two separated phases (Supplementary Fig. 14a), and both
incorporation and adsorption of a Cu atom on are endothermic ( , respectively, see Fig. 5). For , we initially explored the incorporation of Cu in the model with isolated Zn centers (Fig. 5). Cu incorporation ( ) is not favored ( wrt bulk Cu ) even at sites nearby Zn atoms. Interestingly, adsorption of Cu on the surface near Zn sites , Fig. 5) is favored compared to zinc-free , further confirming the affinity. Nevertheless, none of the configurations investigated with separated Zn atoms provide suitable nucleation sites for Cu (Fig. 5). Additionally, none of the models fully agree with the EXAFS and STEM-EDX findings, which indicate that Cu is most likely surrounded by a zinc-rich ensemble under reaction conditions (Supplementary Table 4). However, the formation of CuZn clusters under reaction conditions should be traced back to the affinity, since the formation energy of the bulk (brass) and CuZn alloys are exothermic and , respectively). Hence, alternative scenarios include that the formation of the pairs occurs earlier in the preparation. Indeed, the melting temperatures of and differ significantly (1023 and 2988 K ) meaning that the nucleation during cooling of the FSP could occur independently. This would lead to Zn rich environments with Cu and later the stabilization of these clusters by the oxidic phase ensuring that some O atoms bond the oxophilic Zn ones. Thus, we
Fig. 5 | Theoretical description of copper and zinc sites on the catalyst. Snapshots of DFT models with their associated relative potential energy, with respect to Cu and Zn bulk energies and surface, were employed to rationalize the tendency of Zn and Cu to incorporate into the surface sites and the deposition of Cu on Zn -doped and pure surfaces. The models employed to assess , and catalytic systems via DFT simulations are highlighted in light blue, gray, and purple, respectively. The dashed arrows
represent the incorporation of or Cu atoms (blue and light pink, respectively) accompanied with the formation of oxygen vacancies on the surface. Solid light pink lines denote the deposition of a Cu atom on and surfaces. The common pattern found in and is depicted in black. Color code of DFT models: Zn (blue), Zr (green), Cu (light pink), O (red), and O vacancies (doted white circles).
investigated models where 5 Zn atoms aggregate in nearby sites within the surface with 5 oxygen vacancies. We employed this structure to study the adsorption of a single Cu atom ( ) finding that this is more favored than and (Fig. 5). Including a second Cu atom is exothermic ( ) and results in a CuZn motif very similar to that found in brass, particularly the (110) surface (Fig. 5). Moreover, larger Cu clusters (containing 3 or 4 atoms) are not favored (Supplementary Fig. 16) further underlying the role of nucleation and alloy properties in the final compound. Furthermore, alternative models (over 150 structures) were evaluated to explore the robustness of the ensemble. We assessed different degrees of reduction and the formation of analogous structures with different amount of Zn and Cu , as detailed in the Supplementary Information (see Supplementary Methods and Discussion, Supplementary Fig. 17-21, and Supplementary Table 5). These results indicate that models with 2 copper (possibly forming dimer species) and 5 zinc atoms are more likely thermodynamically and the most compatible with the experimental observations. Therefore, the structure was selected as the most representative model for evaluating the catalytic properties of catalysts from a computational standpoint.

Mechanistic insights into oxygen vacancy formation and dynamics

To investigate the formation and dynamics of oxygen vacancies on Cu , in situ electron paramagnetic resonance spectroscopy (EPR) experiments were carried out on this catalyst and reference and materials (Fig. 6). Fresh shows a weak and almost isotropic peak ( ) and an additional anisotropic signal , Fig. 6a), which are attributed to unpaired electrons trapped in isolated oxygen vacancies (behaving paramagnetically, ) and sites, respectively, as previously reported . Upon thermal treatment in He , the intensity of both signals increased (Fig. 6a), indicating the formation of a relatively small amount of thermally-induced vacancies. In contrast, exposure to the
reaction mixture ( ) led to a decrease in the signal (Fig. 6a and Supplementary Fig. 22a). This is consistent with previous observations indicating that is reoxidized to as electrons are transferred to vacancies ( ) formed by reaction with (see Supplementary Discussion for detailed description) . Still, does not effectively impact the catalytic ensemble properties, acting as a spectator species . Upon switching to reaction mixture a newly broad signal was evidenced and attributed to unpaired electrons delocalized over several oxygen vacancies that interact (exchangecoupled vacancy polarons, (Fig. 6a and Supplementary Fig. 22a). Generally, this broad signal appears when the density of vacancies becomes particularly high, which is considered a key performance descriptor, as reported for systems . For , however, only and species are detected under thermal treatment in He , whereas and signals are not observed (Fig. 6b and Supplementary Fig. 22b) even under , suggesting that a low concentration of vacancies is present on , which could in principle explain its inferior performance (Fig. 3a).
The EPR spectrum of fresh displays a typical signal, with an axial tensor and a well-resolved hyperfine splitting of the component (Fig. 6c). Interestingly, alike , no signal was observed, indicating that sites are better electron scavengers. Thermal treatment in He led to a strong decrease of the and increase of signals, while a new feature appeared around , partially overlapped with the , and consistent with that of surfaceadsorbed superoxide ions (Fig. 6c). However, contrary to , electrons released during these processes are most likely trapped by (i) sites, generating EPR-silent species, (ii) empty vacancies, giving rise to , and (iii) oxygen released from the surface lattice, leading to the formation of the superoxide. This further corroborates its presence significantly promotes vacancy formation. Exposure to the reaction mixture leads to the complete reduction of and to a significant decrease of the signal (Fig. 6c, d). Additionally, a signal assigned to is detected but much broader than the one observed for , indicating that a higher density of vacancies is
Fig. 6 | Oxygen vacancy dynamics during hydrogenation. In situ EPR spectra measured at room temperature of (a) , (b) , and (c) catalysts in fresh form, after heating ramp in He ( , heating rate ) and hydrogenation ( , , and ). (d) Operando EPR spectra of key signals for catalyst under ,
, and dwell time ) and under reaction conditions ( , and ) with continuous time on stream. Evolution of and species, as well as paramagnetic and ferromagnetic oxygen vacancies ( and , respectively) for (e) , (f) , and (g) catalysts under operando conditions described in d and Supplementary Fig. 22a,b.
formed on . The decrease of the signal is most likely linked to its conversion into the broad feature ( ), as the increased concentration of vacancies leads to ferromagnetically-coupled defect sites and, consequently, depletes the number of isolated vacancies ( ).
To gather additional insights into the dynamics of oxygen vacancies on catalyst, operando EPR measurements were performed under He and hydrogenation conditions ( 1 MPa and
573 K; Fig. 6d-g, Supplementary Fig. 22a, b, and 23). Evolution of key species over time (Fig. 6e-g) reveals that and formation rates are mirrored by that of and reduction, suggesting that these processes are linked, thereby validating the above-discussed electron scavenging mechanisms. More importantly, these experiments show that -f signals are better fingerprints compared to for monitoring vacancy formation and annihilation (see Supplementary Discussion for details). Additionally, a comparison of the
reduction kinetics of and shows significant differences (Supplementary Fig. 24), with a monotonous monoexponential decay being observed for . In contrast, clearly shows a component that decays with the same rate as the in , whereas the other starts decreasing at a later stage and is significantly slower, which is most likely related to the reduction of copper nanoparticles present on , in line with microscopy results (Supplementary Fig. 14a). Furthermore, these observations indicate that the overall reduction is more efficient on due likely to a series of coupled factors such as its speciation and strong interaction with zinc and oxygen vacancies. Interestingly, exposing the catalyst to oxygen at room temperature after the reaction reveals a full recovery of the signal (Supplementary Fig. 25), hinting at the stable nature of copper sites. Overall, EPR experiments demonstrate that oxygen vacancy generation is augmented on upon reaction, which most likely contributes to its enhanced performance. Indeed, these findings strongly suggest that these defect sites are an integral component of the low-nuclearity zinc-rich CuZn catalytic ensembles present on , with their role being multifaceted and likely including (i) stabilization of copper atoms and/or reaction intermediates, and (ii) activation of reactants.

Insights on the reaction mechanism

hydrogenation to methanol requires the efficient activation of and , followed by selective proton-hydride transfers to avoid the CO formation through the RWGS reaction. A bifunctional mechanism combining the properties of oxidic ( activation and hydrogenation) and metallic ( activation) phases in close proximity is key to avoid transport of active species and achieve high methanol productivities . Accordingly, adsorption of key reactants, intermediates, and byproduct species ( , and ) on selected models displayed in Fig. 5 was assessed through DFT simulations to rationalize the catalytic properties of , and adsorption is favored on basic sites of whereas activation on this surface is endothermic (Fig. 7), with an activation energy ( ) of 0.70 eV , and occurs heterolytically, forming hydrides and protons ( ) on a pair (Supplementary Fig. 26). In contrast, poorly activates but homolytic dissociation takes place ( ) and the adsorption of its products (hydrogen atoms) is highly favored (Fig. 7). Hence, spillover from Cu nanoparticles onto the surface, where is activated and hydrogenated, is required for the catalyst. Consequently, this leads
to a sluggish transport of active species, which hampers methanol synthesis on this system, thereby contributing to its inferior performance. For , heterolytic activation on a pair is more favored ( and ) compared to (Figs. 7 and 9). Additionally, is activated as carbonate in a nearby site, thus avoiding unnecessary transport of active species onto the catalyst surface, which explains its higher methanol productivity compared to . Adsorption energies for and -dissociated products on the model, which better describes active ensembles on the system, indicate that this configuration enhances activation of these reactants at spatially resolved sites, thereby also circumventing long-range transport of species (Fig. 7). Moreover, activation is barrierless and found to proceed in a homolytic manner on metallic Cu atoms (Figs. 7 and 9), further corroborating the key role of Cu in improving the splitting ability of . Similar results were observed for analog models also comprising partially reduced Zn atoms incorporated into surface lattice positions of , oxygen vacancies, and metallic Cu atoms deposited on a different arrangement ( -b, Supplementary Fig. 27). In contrast, adsorption energies of reactant species on the surface exhibit similar trends as , favored activation and unfavored adsorption, suggesting that partially reduced zinc species, present on , rather than metallic counterparts are required to enable an efficient activation and subsequent transformation of reactants into methanol (Supplementary Fig. 27).
To gain experimental insights into key reactive intermediate species formed on , and catalysts under reaction conditions, operando diffuse reflectance infrared Fourier transform spectroscopy (DRIFTS) experiments were conducted (Fig. 8 and Supplementary Fig. 28). Upon exposure to the reaction mixture ( , and 573 K ; Fig. 8a-c, and Supplementary Fig. 28), all catalysts exhibited bands characteristic of formate ( , , and ) and methoxy species ( , 1140,2827 , and ), which can be assigned to distinct vibrational modes (Supplementary Table 6, 7) . These results demonstrate that methanol synthesis via hydrogenation over all catalysts follows the formate pathway, which is in line with previous studies on and indicates that the copper promotion in Cu is not governed by a different mechanism. Interestingly, a strong signal attributed to gas-phase (Supplementary Fig. 28) is detected for , which is in agreement with its tendency to promote this undesired product, thus explaining its lower
Fig. 7 | Adsorption energies of reactive species. Adsorption energies, of key reactive species on relevant surfaces representative of , and Cu catalytic systems. Color code of DFT models: Zn (blue), Zr (green), Cu (light
pink), and O (red). While only the most energetically favorable of key reactive species for each model is displayed in Fig. 7, other adsorption structures and and are shown in Supplementary Fig. 26 and 27, respectively.
Fig. 8 | Evolution of reaction intermediates. Operando DRIFT spectra of key signals for (a) , (b) , and (c) catalysts under (a-c) hydrogenation ( , and dwell time and (d-f) after switching to a mixture of balanced with
, dwell time ). Evolution of evolution of normalized intensity of formate ( , orange) and methoxy ( C , purple) for (g) , (h) , and (i) catalysts under operando conditions described in (a-f).
compared to and (Fig. 3b). Evolution of and species assessed by monitoring the vibration bands at 2874 and , respectively, over time-on-stream (Fig. ) revealed that their surface concentration changes simultaneously for all catalysts, suggesting that overall reaction rate is determined by the surface concentration. Still, the surface concentration of and on reaches steady state within ca. 30 min , whereas it
requires ca. 50 and 70 min for and , respectively (Fig. ). This suggests that the catalytic ensemble in is more effective in accelerating the formation of these key reaction intermediates. To investigate the dynamics of surface species in pure was replaced by He and the normalized intensity of the vibration bands at 2874 and was continuously monitored (Fig. 8d-i). The concentration of decreases rapidly, gradually,
and slowly for , and , respectively, over 90 min on stream, hinting at a much lower energy barrier for its hydrogenation to on the copper-promoted catalyst (Fig. ). Still, unlike and swiftly hydrogenates newly formed s species, reaching a steady state similar to that in the presence of both and (Fig. 8i). In this case, further hydrogenation of is most likely masked by the dissociative adsorption of methanol on unoccupied sites created by the absence of .
To gain further insights into the reaction mechanism, we computed energy profiles for hydrogenation to methanol and CO on the relevant models presented in Fig. 5 (Supplementary Figs. 29, 30). We followed the formate pathway, which is in line with operando DRIFTS results and previous studies of and other metal oxide-based catalysts . We have also evaluated methanol formation on the surface via CO hydrogenation since gas-phase CO was detected for and systems (Supplementary Fig. 28). Methanol formation is favored over CO production on , , and (Supplementary Fig. 29-30), in agreement with experimental observations (Fig. 3). In contrast, the energy profiles computed on indicate that hydrogenation to methanol competes with CO formation. Still, CO desorption is favored over being further hydrogenated compatible with the lower of Cu . Moreover, DFT simulations indicate that the relative stability between and is favored on over and (Supplementary Table 6). This is in line with the rapid hydrogenation of into methanol observed on compared to and (Fig. ). Overall, based on DFT and DRIFTS, we attribute the high methanol productivity exhibited by to the formation of Zn -rich low-nuclearity CuZn clusters intermixed in the matrix. Specifically, this catalytic ensemble, which is likely stabilized by oxygen vacancies formed upon reaction, efficiently combines both oxidic ( ) and metallic ( Cu ) properties
required to effectively activate and in nearby sites and enable concomitant and fast hydrogenation of key intermediates into methanol (Fig. 9). Finally, given the key role played by copper in promoting , further optimization of its content in the ternary Cu systems should be undertaken in future work.
In summary, copper is uncovered as the most effective promoter for flame-made catalysts for methanol synthesis via hydrogenation, rivaling state-of-the-art palladium promotion, outperforming various hydrogenation metals (i.e., , , or Pt), and leading to a 2 -fold increase in methanol productivity compared to unpromoted . In addition, the catalyst exhibits a significantly more stable and superior performance over 100 h on stream compared to , where copper agglomerates into large nanoparticles, leading to inferior reactivity and methanol selectivity. In contrast, copper forms low-nuclearity metallic species on upon reaction that strongly interact with zinc ensembles intermixed in the surface of the carrier, as evidenced by combining operando XAS analysis and DFT simulations. Operando EPR spectroscopy shows that oxygen vacancy generation is particularly augmented upon reaction and closely associated with the formation of zinc-rich CuZn clusters. The resulting active catalytic ensemble comprising CuZn clusters and oxygen vacancies in the vicinity of Zn atoms efficiently integrate acid-base and splitting functions that enable activation and barrierless homolytic dissociation over spatially resolved sites. In fact, this unique geometric configuration greatly fosters methanol formation through the formate ( ) path, as it promotes fast hydrogenation of to methanol, while preventing undesired long-range transport of active species and CO formation. This work sets an important milestone in the design of metal-promoted catalysts and other systems based on reducible oxides, uncovering the promotional effect of a highly effective architecture based on copper, an earth-abundant element, for hydrogenation to methanol. It also highlights the
Fig. 9 | Comparison of reference and catalysts. Summary of key geometric and reactivity features dictating the performance of catalysts. Snapshots of the adsorption of and the product of dissociative
adsorption on different surfaces are shown for relevant systems. Color code of DFT models: Zn (blue), Zr (green), Cu (light pink), O (red), H (white), and C (dark gray).
potential of stable low-nuclearity metal clusters as a promising class of heterogeneous catalytic materials for diverse related applications.

Methods

Catalyst synthesis

catalysts with a nominal metal promoter and zinc contents of 0.5 and , respectively, were prepared by FSP. Briefly, a precursor solution of dissolved zinc, zirconium, and promoter complexes (see Supplementary Table 1 for a full list) in the desired ratio was pumped through a 0.4 mm nozzle at a flow rate of and dispersed into a fine spray by flowing oxygen at 1.5 bar at a flow rate of . The spray was ignited by a supporting flame generated using of oxygen and of methane. Such particle-generating flames have been well characterized and reported to reach average flame temperatures of , with very fast cooling rates (ca. . The resulting nanoparticles were collected on a glass fiber filter (GF/A-6) and used in hydrogenation without further treatment.

Catalyst characterization

Inductively coupled plasma optical emission spectroscopy (ICP-OES) was carried out using a Horiba Ultra 2 instrument equipped with a photomultiplier tube detector. Samples were digested with the aid of microwave irradiation using a mixture of HCl (Alfa Aesar, ), (Alfa Aesar, ), and HF (Sigma Aldrich, ) with a volume ratio of 2:1:0.5, followed by neutralization with a saturated solution of (Fluka, ). Nitrogen sorption at 77 K was carried out using a Micromeritics TriStar II analyzer. Prior to the measurements, samples were degassed at 473 K under vacuum for 12 h . The specific surface area ( ) was determined using the Brunauer-Emmett-Teller (BET) method. X-ray diffraction (XRD) was conducted using a Rigaku SmartLab diffractometer with a D/teX Ultra 250 detector using Cu Ka radiation ( ) and operating in the Bragg-Brentano geometry. Data was acquired in the range with an angular step size of and a counting time of 1.5 s per step. Temperature-programmed reduction with hydrogen ( -TPR) were measured at ambient pressure using a Micromeritics AutoChem HP II analyzer. The effluent stream of the instrument was measured using thermo-conductivity detector. Highangle annular dark-field scanning transmission electron microscopy (HAADF-STEM) and energy-dispersive X-ray (EDX) spectroscopy was carried out on a Talos F200X equipped with a Super-X EDS system and high-resolution STEM imaging on a probe-corrected HD2700CS (Hitachi), both operated at 200 kV . Catalyst powder was dispersed in ethanol, ground in a mortar, and dispersed on a copper or nickel STEM-grid covered with a lacey carbon support film. After evaporation of ethanol at room temperature, the specimen was mounted on a single-tilt holder and inserted into the microscope. Selected samples were also analyzed by HAADF-STEM-EDX using a probe-corrected Titan Themis operated at 300 kV . Operando X-ray absorption spectroscopy (XAS) experiments were conducted at the Swiss-Norwegian beamlines (SNBL, BM31) of the European Synchrotron Radiation Facility (ESRF) . The X-ray beam was collimated using a double-crystal liquid nitrogen cooled Si(111) monochromator . Metallic Cu and Zn foils were measured for the absolute energy calibration, using -filled ionization chambers for optimal absorption levels. Fluorescence geometry configuration was employed for the analysis under operando conditions, using a oneelement Si SDD detector with Peltier cooling. Continuous scanning was performed for both Cu (between 8880 and 9550 eV ) and Zn edges (between 9550 and 10650 eV ), and the step size was set to 0.5 eV , with a scan duration of 150 s . The beam size was set to (horizontally vertically). The catalyst ( ) was placed between two plugs of quartz wool in a quartz capillary reactor cell ( 1 mm outer diameter, 0.01 mm wall thickness). XAS experiments were carried out in the followings steps: (i) heating to 573 K in , heating rate
, and dwell time ), (ii) switching to hydrogenation conditions ( , dwell time ), (iii) switching to , heating rate , and dwell time 30 min ) and (iv) cooling down to room temperature in , heating rate , and dwell time 30 min ). Extended X-ray absorption fine structure (EXAFS) measurements were performed at each steady state, including the fresh and used sample, as well as under hydrogenation reaction conditions. Under all other conditions, including temperature ramping (steps i and iv) and after switching to different conditions (steps ii and iv, as described above), X-ray absorption near-edge structure (XANES) measurements were instead preferred, to capture the dynamic changes of the systems. The high-purity gases were dosed by a set of Bronckhorst digital mass flow controllers and the outcome was monitored on-line via a Pfeiffer Vacuum OmniStar GSD 320 O mass spectrometer with Quadera software. The resulting spectra were energy calibrated, background corrected, and normalized using the Athena program from the Demeter software suite -weighted EXAFS spectra were fitted in the optimal and -windows (Supplementary Table 5) using the Artemis program. An amplitude reduction factor ( ) of 0.81 and 0.75 were determined by fitting of the EXAFS spectrum of a Cu and Zn foil, respectively. The scattering paths for the fitting were produced using known crystallographic structures of metallic Cu , and , and ZnO . Additional quasi in situ XAS measurements were performed for Re, Ir, and Pt edge, edge, and , and edge. Samples were prepared under the exclusion of air by transferring catalysts samples after catalytic tests ( 20 h under standard conditions, see below) into quartz glass capillaries ( 1 mm outer diameter, 0.01 mm wall thickness, 80 mm length, Hilgenberg GmbH) and sealing of the capillaries in an inert atmosphere. The samples were placed on a stage in place of the setup used for the operando measurements and scanned in fluorescence mode the same Si SDD detector described above. In situ and operando electron paramagnetic resonance (EPR) spectroscopy experiments were performed using a custom-built setup (microwave frequency , center field , sweep width , modulation amplitude , modulation frequency , microwave power , power attenuation , conversion time = 86.55 ms , time constant ). The catalyst sample ( ) was loaded into a quartz capillary ( ) and placed inside an EPR quartz tube (Wilmad; ). The EPR tube was housed at the center of a homemade water-cooled high-temperature resonator , which was installed into a continuous wave (CW) EPR spectrometer (Bruker EMX) operating at X-band frequencies. The reactor was pressurized to 1 MPa , heated in a He flow ( ) to the desired temperature ( ) and kept at this condition for 2 h , while continuously recording EPR spectra. It was subsequently cooled down to room temperature and an EPR spectrum was recorded. The reactor was then heated up again to 573 K and left to stabilize for 30 min . The reaction mixture ( ) was then admitted and was kept flowing ( ) for 2 h . All gases were dosed by a set of digital mass flow controllers and the reactor outlet was monitored online using a Pfeiffer Vacuum Thermo-Star GSD 320 T1 mass spectrometer. The EPR spectra were continuously acquired upon flowing the gases and separately stored, using a 2D acquisition mode, thus enabling a time-resolved monitoring of the process. The reactor was eventually cooled down and EPR spectra were recorded at room temperature. He and air were then subsequently admitted and the EPR spectrum was monitored until no changes were detected. Operando diffuse reflectance infrared Fourier transform spectroscopy (DRIFTS) was conducted in a custom-built setup comprising a gas dosing unit, back-pressure regulators, a Pfeiffer Vacuum OmniStar GSD mass spectrometer for online product analysis, and a Bruker INVENIO-S spectrometer equipped with a liquid nitrogen cooled mercury cadmium telluride (MCT) detector. In a typical experiment, 50 mg of the undiluted sample was packed into a high-
temperature reaction chamber with ZnSe windows (Harrick Scientific), which was mounted in a Praying Mantis diffuse reflection attachment (Harrick Scientific). After purging the sample with flowing , PanGas, 4.6) for 60 min , the reaction chamber was pressurized to 1.5 MPa and heated to 573 K with a ramp of . Subsequently, spectra were continuously acquired at a spectral resolution of by accumulating 128 scans every 30 s and were normalized against a KBr background taken at 573 K and 1.5 MPa . The gas feed was then switched to a mixture of (PanGas, 5.0) and in , Messer, 4.5) with a molar ratio of 4 at a total flow rate of . After 90 min under reaction conditions, the flow was stopped and ) balanced with was fed into the reaction chamber for an additional 90 min . The raw data obtained with Bruker OPUS 8.2 software were processed and difference spectra as well as intensity profiles were generated with a self-coded Python routine, available as open-source on GitHub (https://github.com/philpreikschas/ operando-ir).

Catalyst evaluation

The gas-phase hydrogenation of to methanol was conducted using a PID Eng&Tech high-pressure continuous-flow setup comprising four parallel fixed-bed reactors, as described elsewhere . Undiluted catalysts (mass, ; particle size ) were loaded into each reactor tube (internal diameter 4 mm ), held in place by a quartz-wool bed set on a quartz frit, and purged in flowing He , PanGas, 4.6) for 30 min . Under the same flow, the pressure was increased to 5 MPa for a leak test, which was followed by heating up the catalyst bed to . The reaction was carried out by feeding a mixture of (PanGas, 5.0 ), in , Messer, 4.5), with a molar ratio of 4 at , and a gas hourly space velocity (GHSV) of , unless stated otherwise. The selectivity of the catalysts was compared at a constant degree of conversion ( ) as described in Fig. 3b by adjusting the GHSV for each system.
The effluent streams were analyzed by gas chromatography every 1 h . Response factors ( ) for each compound , respective to the internal standard ( in He , Messer, purity 3.5), in the GC analysis were determined by Eq. (1):
where is the integrated area determined for the peak of compound and is the corresponding known molar flow at the reactor inlet. An average of 5 points around the expected analyte concentration was used. The unknown effluent molar flow of compound was determined using Eq. (2):
where is the known flow of at the reactor outlet.
Conversion ( ), selectivity ( ), and methanol production rate ( МеОН ) were calculated using Eqs. (3-5):
The methanol space-time yield ( ) is the product of and the molar weight ( ) of methanol ( ) as stated in Eq. (6):
The carbon balance was determined for each experiment according to Eq. (7):
and was always within a margin.

Computational details

Density function theory (DFT) simulations were conducted with Vienna ab initio simulation package (VASP) and the Perdew-BurkeErnzerhof (PBE) density functional . Valence electrons were expanded from a plane-wave basis set with a kinetic cut-off energy of 500 eV while core electrons were described with projector augmented-wave (PAW) pseudopotentials . The Brillouin zone was sampled with a -centered mesh with a reciprocal grid size narrower than , which was obtained with the Monkhorst-Pack method . Lattice parameters for , and were optimized with a kinetic energy cut-off of at least 700 eV . Then, slab models for , and terminations were cleaved from the optimized bulks and represented with 2 ), , and with expansions, respectively. In all cases, the two bottommost layers were kept fixed and the two outermost were allowed to relax. We added a vacuum region of between slabs and a dipole correction along the axes was applied to account for the asymmetry in the relaxations . The incorporation of Zn and Cu , as well as the formation of oxygen vacancies were assessed at different sites of following the procedure employed in our previous work . Additionally, the deposition of single atoms and small clusters (up to 4 atoms) of Cu was explored in models of pure and Zn -doped surfaces. The adsorption energies of key molecules ( , , and ) was carried out in relevant surfaces to rationalize the catalytic properties of , and . Clean surfaces and gas-phase , and CO molecules were employed as thermodynamic sinks. The adsorptions of and were performed with the adsorption of an H atom in nearby sites as OH or to avoid the adsorption of fragments with unpaired electrons that modify the electronic properties of semiconducting metal-oxides . Furthermore, the effect of weak long-range interactions and the selfinteraction error on the adsorption energies were evaluated (Supplementary Tables 8 and 9) . Dispersion corrections were included via Grimme’s D3 approach while a Hubbard correction by means of the Dudarev approach with a value of 4.00 eV was applied on the states of Zr atoms .
Energy profiles for hydrogenation to methanol and the competitive RWGS to form CO were computed on selected models of the , and catalysts. Clean surfaces and gasphase , and molecules were used as thermodynamic sinks. The climbing image nudged elastic band (CI-NEB ) was employed to locate transition states. Their nature was confirmed by computing numerical frequencies with a step size of .

Data availability

Data presented in the main figures of the manuscript are publicly available through the Zenodo repository (entry: 8309938, link: https:// doi.org/10.5281/zenodo.8309938) . Inputs and outputs for all DFT simulations can be found online in the ioChem-BD repository , at https://iochem-bd.iciq.es/browse/review-collection/100/64621/ fd52044500da8d298dc081fd. Further data supporting the findings of this study are available in the Supplementary Information and source
data. All other relevant source data are available from the corresponding author upon request. Source data are provided with this paper.

Code availability

Source data of the Python application used for processing and analyzing operando infrared spectroscopy data are open-sourced on GitHub (https://github.com/philpreikschas/operando-ir) and additionally available in the Zenodo repository (https://doi.org/10.5281/ zenodo.10818472)

References

  1. Li, K. & Chen, J. G. Hydrogenation to methanol over containing catalysts: insights into induced synergy. ACS Catal. 9, 7840-7861 (2019).
  2. Wang, J. et al. A highly selective and stable solid solution catalyst for hydrogenation to methanol. Sci. Adv. 3, 1-11 (2017).
  3. Han, Z. et al. hydrogenation to methanol on solid solution catalysts with ordered mesoporous structure. J. Catal. 396, 242-250 (2021).
  4. Pinheiro Araújo, T. et al. Design of flame-made catalysts for sustainable methanol synthesis from . Adv. Energy Mater. 13, 2204122 (2023).
  5. Pinheiro Araújo, T. et al. Methanol synthesis by hydrogenation of hybrid feeds. ChemSusChem 14, 2914-2923 (2021).
  6. Lee, K. et al. Atomic Pd-promoted solid solution catalyst for hydrogenation to methanol. Appl. Catal. B: Environ. 304, 120994 (2022).
  7. , Hong, X. & Liu, G. Highly dispersed metal doping to ZnZr oxide catalyst for hydrogenation to methanol: Insight into hydrogen spillover. J. Catal. 393, 207-214 (2021).
  8. Huang, C. et al. hydrogenation to methanol over PdZnZr solid solution: effects of the PdZn alloy and oxygen vacancy. ACS Appl. Energy Mater. 4, 9258-9266 (2021).
  9. Lee, K. et al. Engineering nanoscale H supply chain to accelerate methanol synthesis on . Nat. Commun. 14, 819 (2023).
  10. Pinheiro Araújo, T. et al. Flame-made ternary catalyst with enhanced oxygen vacancy generation for hydrogenation to methanol. Nat. Commun. 13, 5610 (2022).
  11. Pinheiro Araújo, T. et al. Flame spray pyrolysis as a synthesis platform to assess metal promotion in -catalyzed hydrogenation. Adv. Energy Mater. 12, 2103707 (2022).
  12. Wang, J. et al. Hydrogenation to methanol over -based catalysts: from mechanism to catalyst development. ACS Catal. 11, 1406-1423 (2021).
  13. Frei, M. S. et al. Atomic-scale engineering of indium oxide promotion by palladium for methanol production via hydrogenation. Nat. Commun. 10, 1-11 (2019).
  14. Bavykina, A. et al. Turning a methanation Co catalyst into an In-Co methanol producer. ACS Catal. 9, 6910-6918 (2019).
  15. Rui, N. et al. Hydrogenation of to methanol on a catalyst. ACS Catal. 10, 11307-11317 (2020).
  16. Snider, J. L. et al. Revealing the synergy between oxide and alloy phases on the performance of bimetallic In-Pd catalysts for hydrogenation to methanol. ACS Catal. 9, 3399-3412 (2019).
  17. Frei, M. S. et al. Nanostructure of nickel-promoted indium oxide catalysts drives selectivity in hydrogenation. Nat. Commun. 12, 1960 (2021).
  18. Kaiser, S. K. et al. Performance descriptors of nanostructured metal catalysts for acetylene hydrochlorination. Nat. Nanotechnol. 17, 606-612 (2022).
  19. Pinheiro Araújo, T. et al. Reaction-induced metal-metal oxide interactions in Pd- catalysts drive selective and Stable hydrogenation to methanol. Angew. Chem. Int. Ed. 62, 202306563 (2023).
  20. Beck, A. et al. Following the structure of copper-zinc-alumina across the pressure gap in carbon dioxide hydrogenation. Nat. Catal. 4, 488-497 (2021).
  21. Tsoukalou, A. et al. Operando X-ray absorption spectroscopy identifies a monoclinic :In solid solution as the active phase for the hydrogenation of to methanol. ACS Catal. 10, 10060-10067 (2020).
  22. Zabilskiy, M. et al. Mechanistic study of carbon dioxide hydrogenation over Pd/ZnO-based catalysts: the role of palladium-zinc alloy in selective methanol synthesis. Angew. Chem. Int. Ed. 60, 17053-17059 (2021).
  23. Prins, R. Hydrogen spillover: facts and fiction. Chem. Rev. 112, 2714-2738 (2012).
  24. Bahruji, H. et al. catalysts for direct hydrogenation to methanol. J. Catal. 343, 133-146 (2016).
  25. Zhao, H. et al. The role of species in single-atom catalyst for hydrogenation. Nat. Catal. 5, 818-831 (2022).
  26. Tada, S. et al. Ag addition to CuO-ZrO 2 catalysts promotes methanol synthesis via hydrogenation. J. Catal. 351, 107-118 (2017).
  27. Samson, K. et al. Influence of structure and copper electronic state on activity of catalysts in methanol synthesis from . ACS Catal. 4, 3730-3741 (2014).
  28. Ma, Y. et al. Reactivity of a zirconia-copper inverse catalyst for hydrogenation. J. Phys. Chem. C 124, 22158-22172 (2020).
  29. Wang, Y. et al. Strong evidence of the role of in affecting methanol selectivity from hydrogenation over . Chem 6, 419-430 (2020).
  30. Yang, C. et al. Strong electronic oxide-support interaction over for highly selective hydrogenation to methanol. J. Am. Chem. Soc. 142, 19523-19531 (2020).
  31. Wu, C. et al. Inverse as a highly efficient methanol synthesis catalyst from hydrogenation. Nat. Commun. 11, 5767 (2020).
  32. . et al. Insights into the interfacial structure of catalysts for methanol synthesis from hydrogenation: Effects of Cusupported nano-ZrO inverse interface. Chem. Eng. J. 470, 144006 (2023).
  33. Frei, M. S. et al. Mechanism and microkinetics of methanol synthesis via hydrogenation on indium oxide. J. Catal. 361, 313-321 (2018).
  34. Mädler, L. & Pratsinis, S. E. Bismuth oxide nanoparticles by flame spray pyrolysis. J. Am. Ceram. Soc. 85, 1713-1718 (2004).
  35. Kemmler, J. A. et al. Quenched, nanocrystalline high temperature phase for gas sensing applications. Sens. Actuators B: Chem. 161, 740-747 (2012).
  36. Meierhofer, F., Mädler, L. & Fritsching, U. Nanoparticle evolution in flame spray pyrolysis: process design via experimental and computational analysis. AIChE J. 66, 1-14 (2020).
  37. van Beek, W., Safonova, O. V., Wiker, G. & Emerich, H. SNBL, a dedicated beamline for combined in situ X-ray diffraction, X-ray absorption and Raman scattering experiments. Phase Transit. 84, 726-732 (2011).
  38. Ravel, B. & Newville, M. ATHENA, ARTEMIS, HEPHAESTUS: data analysis for X-ray absorption spectroscopy using IFEFFIT. J. Synchrotron Radiat. 12, 537-541 (2005).
  39. Zichittella, G., Polyhach, Y., Tschaggelar, R., Jeschke, G. & PérezRamírez, J. Quantification of redox sites during catalytic propane oxychlorination by operando EPR spectroscopy. Angew. Chem. Int. Ed. 60, 3596-3602 (2021).
  40. Pinheiro Araújo, T. et al. Impact of hybrid feeds on methanol synthesis over -based catalysts. Appl. Catal. B: Environ. 285, 119878 (2021).
  41. Kresse, G. & Furthmüller, J. Efficiency of ab-initio total energy calculations for metals and semiconductors using a plane-wave basis set. Comput. Mater. Sci. 6, 15-50 (1996).
  42. Kresse, G. & Furthmüller, J. Efficient iterative schemes for ab-initio total-energy calculations using a plane-wave basis set. Phys. Rev. B 54, 11169-11186 (1996).
  43. Perdew, J. P., Burke, K. & Ernzerhof, M. Generalized gradient approximation made simple. Phys. Rev. Lett. 77, 3865-3868 (1996).
  44. Blöchl, P. E. Projector augmented-wave method. Phys. Rev. B 50, 17953-17979 (1994).
  45. Kresse, G. & Joubert, D. From ultrasoft pseudopotentials to the projector augmented-wave method. Phys. Rev. B 59, 1758-1775 (1999).
  46. Monkhorst, H. J. & Pack, J. D. Special points for Brillouin-zone integrations. Phys. Rev. B 13, 5188-5192 (1976).
  47. Ricca, C., Ringuedé, A., Cassir, M., Adamo, C. & Labat, F. A comprehensive DFT investigation of bulk and low-index surfaces of polymorphs. J. Comput. Chem. 36, 9-21 (2015).
  48. Christensen, A. & Carter, E. A. First-principles study of the surfaces of zirconia. Phys. Rev. B 58, 8050-8064 (1998).
  49. Mizutani, U., Takeuchi, T. & Sato, H. Interpretation of the HumeRothery rule in complex electron compounds: -phase Alloy, FK-type and MI-type approximants. Prog. Mater. Sci. 49, 227-261 (2004).
  50. Jain, A. et al. Commentary: The materials project: a materials genome approach to accelerating materials innovation. APL Mater. 1, 011002 (2013).
  51. Makov, G. & Payne, M. C. Periodic boundary conditions in ab initio calculations. Phys. Rev. B 51, 4014-4022 (1995).
  52. Castelli, I. E. et al. Role of the band gap for the interaction energy of coadsorbed fragments. J. Phys. Chem. C 121, 18608-18614 (2017).
  53. Grimme, S., Ehrlich, S. & Goerigk, L. Effect of the damping function in dispersion corrected density functional theory. J. Comput Chem. 32, 1456-1465 (2011).
  54. Hubbard, J. Electron correlations in narrow energy bands II. the degenerate band case. Proc. R. Soc. A: Math. Phys. Eng. Sci. 277, 237-259 (1963).
  55. Dudarev, S. L., Botton, G. A., Savrasov, S. Y., Humphreys, C. J. & Sutton, A. P. Electron-energy-loss spectra and the structural stability of nickel oxide: An LSDA + U study. Phys. Rev. B 57, 1505-1509 (1998).
  56. Koga, H. et al. Facile NO-CO elimination over zirconia-coated 0) surfaces: Further evidence from DFT + U calculations. Appl Surf. Sci. 508, 145252 (2020).
  57. Syzgantseva, O. A., Calatayud, M. & Minot, C. Revealing the surface reactivity of zirconia by periodic DFT calculations. J. Phys. Chem. C 116, 6636-6644 (2012).
  58. Ruiz Puigdollers, A., Illas, F. & Pacchioni, G. Effect of nanostructuring on the reactivity of zirconia: a DFT + U study of au atom adsorption. J. Phys. Chem. C 120, 17604-17612 (2016).
  59. Tang, X., Zhang, H., Sun, C., Qiao, X. & Ju, D. Adsorption mechanisms over -modified surface for and CO): a DFT + U study. Surf. Sci. 716, 121976 (2022).
  60. Henkelman, G., Uberuaga, B. P. & Jónsson, H. A climbing image nudged elastic band method for finding saddle points and minimum energy paths. J. Chem. Phys. 113, 9901-9904 (2000).
  61. Pinheiro Araújo, T. et al. In situ formation of low-nuclearity CuZn ensembles on catalysts promotes hydrogenation to methanol [Data set]. Zenodo https://doi.org/10.5281/zenodo. 8309938 (2023).
  62. Álvarez-Moreno, M. et al. Managing the computational chemistry big data problem: the ioChem-BD platform. J. Chem. Inf. Model. 55, 95-103 (2015).
  63. Bo, C., Maseras, F. & López, N. The role of computational results databases in accelerating the discovery of catalysts. Nat. Catal. 1, 809-810 (2018).
  64. Preikschas, P. In situ formation of low-nuclearity CuZn ensembles on catalysts promotes hydrogenation to methanol [Code – philpreikschas/operando-ir: v0.1.1 (v0.1.1)]. Zenodo https:// doi.org/10.5281/zenodo. 10818472 (2024).

Acknowledgements

This publication was created as part of NCCR Catalysis (grant number 180544), a National Centre of Competence in Research funded by the Swiss National Science Foundation. The Scientific Center for Optical and Electron Microscopy (ScopeM) at ETH Zurich is thanked for access to their facilities. Mr. Dario Faust Akl and Mr. Henrik Eliasson are thanked for acquiring the HAADF-STEM-EDX data. T.Z. thanks the Agency for Science, Technology and Research (A*STAR) Singapore for support through a graduate fellowship. The Spanish Ministry of Science and Innovation is acknowledged for financial support (PRE2019-088791, PID2021-122516OB-IOO, and Severo Ochoa Grant MCIN/AEI/10.13039/ 501100011033 CEX2019-000925-S) and the Barcelona Supercomputing Center-MareNostrum (BSC-RES) for providing generous computer resources. Pol Sanz Berman is thanked for his support with graphic design. The Swiss Norwegian beamlines (SNBL, ESRF) are acknowledged for the provision of beamtime and its staff for invaluable support.

Author contributions

T.P.A. and J.P.-R. conceived and coordinated all stages of this research. T.P.A., G.G., J.M.V, N.L., S.M., and J.P.-R. wrote the article with input from all other co-authors. T.P.A. and Z.R.B. conducted catalyst characterization and evaluation. J.M.V and N.L. conducted the density functional theory simulations. F.K. conducted the electron microscopy analyses. M.A. and G.J. conducted the electron paramagnetic resonance spectroscopy studies. G.G. evaluated X-ray absorption spectroscopy data. P.O.W., R.N.G., and W.J.S. prepared the catalysts using the flame spray pyrolysis method. T.Z. conducted kinetic analyses. P.P. supervised acquisition and evaluated diffuse reflectance infrared Fourier transform spectroscopy data. J.P.-R. supervised the entire project and managed resources and funding. All the authors provided inputs to the manuscript and approved the final version.

Competing interests

The authors declare no competing interests.

Additional information

Supplementary information The online version contains
supplementary material available at
https://doi.org/10.1038/s41467-024-47447-6.
Correspondence and requests for materials should be addressed to Javier Pérez-Ramírez.
Peer review information Nature Communications thanks the anonymous reviewers for their contribution to the peer review of this work. A peer review file is available.
Reprints and permissions information is available at http://www.nature.com/reprints
Publisher’s note Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.
Open Access This article is licensed under a Creative Commons Attribution 4.0 International License, which permits use, sharing, adaptation, distribution and reproduction in any medium or format, as long as you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons licence, and indicate if changes were made. The images or other third party material in this article are included in the article’s Creative Commons licence, unless indicated otherwise in a credit line to the material. If material is not included in the article’s Creative Commons licence and your intended use is not permitted by statutory regulation or exceeds the permitted use, you will need to obtain permission directly from the copyright holder. To view a copy of this licence, visit http://creativecommons.org/ licenses/by/4.0/.
© The Author(s) 2024

  1. Institute of Chemical and Bioengineering, Department of Chemistry and Applied Biosciences, ETH Zurich, Vladimir-Prelog-Weg 1, 8093 Zurich, Switzerland.
    Institute of Chemical Research of Catalonia (ICIQ-CERCA), The Barcelona Institute of Science and Technology, Av. Països Catalans 16, 43007
    Tarragona, Spain. Laboratory of Physical Chemistry, Department of Chemistry and Applied Biosciences, ETH Zurich, Vladimir-Prelog-Weg 2, 8093 Zurich, Switzerland. Department of Inorganic Chemistry and Materials Institute (IUMA), Faculty of Sciences, University of Alicante, Ap. 99, E-03080 Alicante, Spain. Laboratory of Inorganic Chemistry, Department of Chemistry and Applied Biosciences, ETH Zurich, Vladimir-Prelog-Weg 1, 8093 Zurich, Switzerland. e-mail: jpr@chem.ethz.ch