Discovery and Comparison of Homogeneous Catalysts in a Standardized HOT‐CAT Screen with Microwave‐Heating and qNMR Analysis: Exploring Catalytic Hydration of Alkynes

A HOT‐CAT (homogeneous thermal catalysis) screen using microwave‐heating and quantitative NMR (qNMR) analysis has been developed for identification and comparison of catalyst activity in homogeneous metal‐based catalysis. The hydration of terminal alkynes to ketones or aldehydes served as a model reaction in this proof‐of‐concept study. Key aspects of the screen are the use of a high‐temperature setting (e. g., 160 °C) at a fixed, short reaction time (e. g., 15 min) for all samples. Analysis of crude reaction mixtures by a standardized, quantitative 1H NMR protocol gives a comprehensive picture of catalyst chemo‐ and regioselectivity, which permits broad comparisons and the discovery of non‐target reactivity. For catalytic alkyne hydration, data for 105 runs involving 81 catalyst systems with 15 different metals is presented. The activity of all established catalyst systems was reproduced, and new catalyst systems with Markovnikov hydration selectivity were discovered and applied to preparative runs, namely Cu2O−CSA (CSA=camphorsulfonic acid), Co(OAc)2−tetraphenylporphyrin−CSA and [IrCl(COD)]−CSA.


Introduction
The identification of new catalysts for synthetic transformations is important for progress in synthetic organic chemistry. [1] Any new catalytic system that displays unique activity and selectivity patterns may help solving a specific synthetic problem. In searching new catalysts for an organic transformation one is first faced with choosing reaction conditions by defining relevant parameters for variables like temperature, pressure, solvent, reactant ratios, concentration and reaction time. [2] Limiting the discussion to homogeneous, metal-complex-catalyzed reactions, one will subsequently generate a potential catalyst by combining a metal precursor, steering ligand and various co-catalytic additives. [3] Such substance variables and their relative settings (ratios) define a catalyst system, whereas catalyst loading is rather a reaction condition. [4] Considering the large number of independent variables and parameter settings, the de novo search for a catalyzed reaction can be inherently complex, even for a predefined target transformation. [2a,5] Over the past years we have observed the development process for homogeneous catalytic hydration of alkynes to carbonyl compounds for application in organic synthesis. [6,7] In new research papers on alkyne hydration catalysts, there often is a common structure: [8] (I) A potential new catalyst system is proposed based on a mechanistic hypothesis or by analogy to established catalysts. (II) A screening with a suitable assay is performed, i. e. a suitable model reaction, whose product is readily detected, is performed in the presence of the potential catalyst. (III) Catalysis is confirmed through analysis or isolation of a target product in quantities exceeding those found in blank reactions. (IV) Key parameters of the single most promising catalyst system are optimized in a focused screen, until the model reaction reaches a satisfactory product yield. (V) The new catalyst system is applied to a selection of diversely cofunctionalized substrates, with parameter settings optimized for the model substrate, and product yields are recorded as (exclusive) indicators of efficiency.
Certain aspects of this operation mode seem unsatisfactory: a) Restricting a study to a single type of catalyst system avoids recognizing interdependencies between chemically different, but related catalyst systems; b) optimizing a reaction to the maximal yield of a model target, while disregarding the nature and amount of side-products formed, ignores selectivity aspects and limits the predictive value of optimization data; c) the common approach to run a reaction to consumption of the starting material (at variable reaction times), or the failure to quantify unreacted starting material (at fixed reaction times), render comparisons among catalyst systems difficult, because the target yield intertwines activity and selectivity aspects ( Figure 1).
We now present a case-study in catalyst-screening that aims at collecting comprehensive information on catalyst activity and selectivity for a specific model reaction. Widely differing catalyst systems will be compared by activity and selectivity, based on a standardized experimental procedure. The kind of information we want to obtain is that usually found in a review article, where it is normally compiled from various individual studies and where direct comparisons are not possible, since all studies covered tend to use different model reactions and reaction conditions.
The reaction in case is catalytic hydration of terminal alkynes to aldehydes or ketones (Scheme 1). [6] This transformation is of synthetic interest, involves a fundamental selectivity (Markovnikov vs. anti-Markovnikov regioselectivity) and has already been catalyzed by a number of metal-and other element-based species. Our 2007 review listed catalytic activities for Brønsted acids, enzymes, Hg, other metals (Ce, W; Fe; Ru, Rh, Ir, Pd, Pt, Cu, Ag, Au, Zn, Cd, Tl) and a half-metal (Te). In total, hydration activity was listed for catalysts based on ca. 17 elements. [6] Over the past 12 years, alkyne hydration activity has been reported for additional elements (Ca, [9] Co, [10,11] Ga, [8c] In, [12] Tm, [8d] Yb, [8c] Sc, [8c] Bi, [8a] Y, [8a] Eu, [8a] La, [8a] Sn [8e] ) while one reported activity has been refuted (W). [13] Many more new catalyst systems have been described for elements with previously established activity, the frequency of reports following the order Au @ Cu, Ag > Pt, Ru, Fe, Pd. We now wish to explore a generalized, systematic activity screening that can be applied to any type of potential alkyne hydration catalysts, and which will provide information on the relative activity and selectivity of each catalyst. As a secondary aim, we hope to find new and practically useful catalyst systems for alkyne hydration with either Markovnikov or anti-Markovnikov selectivity.

Results
Definition of the screening procedure and conditions. The reaction temperature in a comparative catalyst screening can be set to the lowest value for achieving notable, yet incomplete conversions. Relative catalyst activities are then derived by comparing the yield of target product after a fixed reaction time. [14] If a highly active catalyst emerges from the screening, the temperature setting may be lowered or the catalyst loading reduced for further optimization. [14,15] We have defined a different HOT-CAT (homogeneous thermal catalysis) approach for catalysis screening, where catalytic reactions are screened in a microwave reactor at deliberately high temperature settings. [16,17] Reaction times are fixed at short duration (15-30 min) to allow for serial experiments in a mono-mode reactor with auto-sampler. A boundary condition of our screen is that the desired catalytic activity will be scarce. Detecting the desired activity at all is a priority, no matter if it is high or low. An inherent assumption of our approach is that low activity catalysts will be easier to detect at a very high temperature setting. More active catalysts tend to show activity well above their usual temperature optimum and will therefore also be detected in the screen. We propose that the HOT-CAT approach is most suited to detect catalytic activity with the least number of (e. g., a single) experiments. The second key aspect of our Scheme 1. General scheme of a catalyzed transformation (alkyne hydration). The reaction system is defined by setting parameters for general reaction conditions (variables), and for the subset of variables that make up the catalyst system. Each parameter setting affects the reaction output, measured as product composition and expressed in derived quantities.
approach is quantification of all relevant reaction components to define catalyst selectivity. For efficiency, analysis should be limited to a single measurement, which places restrictions on the model reaction and suitable analytical techniques.
In short, we wish to perform a fast, efficient screening for new catalytic reactivity under HOT-CAT conditions in a microwave-reactor, use a deliberately high reaction temperature at a fixed, short reaction time and gain comparative information on a large number of catalyst systems, their relative activity, chemo-and stereoselectivity, through quantitative product analysis. Questions to be answered are: is the HOT-CAT approach viable for new catalyst discovery? Does it reproduce known catalytic activity under screening conditions to validate the approach? Can we detect new catalysts for a defined model reaction? Will unexpected new catalytic activity towards unpredicted products be found?
Screening System 1: Microwave heating/GC-MS analysis. Our first approach to catalyst screening centered on the hydration of 1-heptyne in aqueous acetone. As secondary substrate, 1-octene was added to the test mixture with the intention to concomitantly screen for catalytic alkene hydration. [17,18] This part of the project proved futile and will not be further discussed. Screening experiments were run under the conditions defined in Scheme 2. Di-n-butyl ether was added to the cooled reaction mixture as internal standard for GC-MS analysis. [19] Hydration products 2-heptanone, 1-heptanal and remaining 1-heptyne were quantified through calibration with reference samples.
This screening system detects catalytic hydration activity at a threshold of ca. 1 mol% of either Markovnikov-(2-heptanone) or anti-Markovnikov-(1-heptanal) product formation. The GC-MS-analysis was performed directly from the reaction mixture. Disadvantages of this approach, besides the not overly high sensitivity, were low substance recoveries, which implied that unknown reaction products had escaped detection. The screening was performed on a total of ca. 60 potential metallic catalysts which were either tested as such, in combination with an ambifunctional pyridylphosphane ligand, [20,21] and other additives (CSA, AgOTf). The results from 159 runs thus performed are listed in the supporting information (Table S1) and will be referred to in the ensuing text where adequate. Eventually the first screening system was abandoned because of the difficulty of obtaining accurate quantitative data and satisfactory recoveries.
Screening System 2: Microwave heating/qNMR analysis. Undec-10-yn-1-ol (1) is commercially available as liquid alkyne that is easily purified by vacuum distillation and can be dosed by microliter syringe. The compound and its follow-up-products are little volatile, have low water solubility and are readily recovered by extractive workup. The hydration products ketone 2 (δ 2.12, 3 H) and aldehyde 3 (δ 9.77, 1 H) display characteristic, isolated peaks suitable for quantitative analysis by 1 H NMR spectroscopy. Unreacted 1 is detected through its acetylenic CÀ H (δ 1.94), while the R-CH 2 OH signal (δ 3.64, 2 H) provides the total of products derived from 1 with intact alcohol endgroups. Multiple non-target products were also detected (Scheme 3; for qNMR diagnostics, see Table S2): acetals (4, 5) result from acetalization of 2/3 and alcoholic solvent (MeOH, iPrOH), or via direct addition of the latter to alkyne. Allene 6 was an impurity in commercial 1, but else was not significantly formed in the catalytic runs. Through hydride transfer and isomerization reactions, alkyne 1 is transformed into a hydrocarbon chain (7), a terminal alkene (8), or into isomeric internal (9) alkenes. [22] Transfer-hydrogenation of aldehyde 3 to diol 10 as secondary reaction was sometimes seen with iPrOH as solvent. Generation of 10 raises [CH 2 OH] above 100 mol% relative to initial 1, permitting an approximate quantification. [23] Alkyne trimerization [24] returns arenes 11 a/b, whereas alkyne dimerization [25] potentially generates (E/Z)-stereoisomers of regioisomers 12 a/b. Recovery is the sum of all individually detected components 1-12 in mol %. In case it amounts to less than [CH 2 OH], there must be undetected components derived from 1 in the sample (vide infra). The difference is listed as unknown and serves as control for the integrity of analysis.
In practice, the screening procedure consisted in combining a potential catalyst with degassed solvent, water and 1 in a microwave tube, followed by heating to 160°C for 15 minutes. After addition of internal standard and following a suitable workup-procedure, the sample was analyzed by quantitative 1 H NMR spectroscopy. Catalytic reactions were performed in one or several solvents (A: iPrOH, B: acetone, C: methanol, D: NMP) with water (4 : 1 (v/v)). Such comparatively water-rich, non-acidic media were deliberately chosen to reduce unspecific Brønsted acid catalysis as opposed to metal-specific reactivity. [26] Validation of the screening approach with established alkyne hydration catalyst systems. Initial experiments focused on established catalyst systems for anti-Markovnikov hydration of terminal alkynes to check if their reported activities and selectivity are reproduced under HOT-CAT screening conditions (Table 1). Wakatsuki's catalyst CpRuCl(dppm) [27] (dppm = 1,2-bisdiphenylphosphino-methane) gave close to 90 % aldehyde 3 at a regioselectivity of > 200 : 1 (entry 1). The conversion was incomplete at lower catalyst loading, due to deactivation as confirmed by detection of 1-decanol, formed via R-CH 2 -CH 2 -CO[Ru]. [28] The reported activity ("reactions using 2-10 mol % of catalyst in 2-propanol at 100°C gave the desired aldehydes in good to excellent yields after 12 h") is well mirrored by the 15 min HOT-CAT experiment. A combination of CpRu + -precursor complex [CpRu(C 10 H 8 )]PF 6 [29] with dppm presented comparably high activity (entry 3), whereas the lower activity of a CpRu + -dppe system (entry 4) mirrors Wakatsuki's findings with the latter ligand. [27] The choice of precursor is critical since CpRuCl (PPh 3 ) 2 with dppm is unreactive (entry 5); inactive [CpRu(η 2dppm)(PPh 3 )] + may have been formed (cf. Table S1, entry 93). The catalyst system CpRuCl(PPh 3 ) 2 À ISIPHOS (ISIPHOS = 2-(diphenylphosphino)-6-(2,4,6-triisopropylphenyl)pyridine) [21,30] performs predictably well (entries 6, 7), with entry 6 mirroring our previously published HOT-CAT conditions. [30] Although entries 1-6 use optimal solvent mixtures for the respective catalyst systems, catalytic activity is also traced in a non-standard NMPÀ H 2 O medium (entry 7). In practice, if novel but low activity was observed in one medium, we repeated the experiment in other media.
For the first time, catalytic activity was detected with osmium complexes (entries 12-15). [6] Hydration with hexachloroosmate(IV) is Markovnikov-selective, and most notable in methanol (entry 15). The overall low chemoselectivity is due to a preference of transfer hydrogenation and alkene isomerization reactions of this catalyst. The large proportion of "unknowns" points to alkyne polymerization as other sidereaction (vide infra).
Group 9 metals. Naka found catalytic activity of watersoluble, sulfonated cobalt(III)porphyrins for Markovnikov hydration of terminal alkynes under relatively mild conditions (MeOHÀ H 2 O, 80°C, 6-16 h; neutral or slightly acidic). [10] Cationic cobalt(III)salen complexes were also found to hydrate arylalkynes in acidic media, but less so aliphatic alkynes. [11] The simple salts CoCl 2 and Co(OAc) 2 were inactive in the 1-heptyne screen (Table S1, entries 101-105). Cyanocobalamin (vitamin B 12 ) was included in both screens, but showed no catalytic activity, neither in presence of acid nor with iron(III)triflate as potential cyanide scavenger (Table S1, entries 106, 107; Table 4, entries 1-4). An in situ Co-porphyrin catalyst from cobalt(II) acetate and TPP (tetraphenylporphyrin) under argon displayed Markovnikov hydration activity after addition of CSA (entry 6); higher conversions were obtained in air (entry 7) and even more so with excess acid (entry 8). The active species in this system is likely similar to the Co(III)-porphyrin of Naka, [10] but is based on the accessible ligand TPP, [50] and thus was selected for preparative evaluation (vide infra). No activity was observed with Co(II)salen complexes -at least in the absence of Brønsted acid (entries 9, 10). Cobalt cyclobutadiene sandwich complex, [(C 4 Me 4 )Co(C 6 H 6 )]BF 4 (15) [49c] (Figure 2) was inactive either alone or with ambifunctional steering ligands (entries 11-13).
Preparative runs with new catalyst systems. The utility of reaction conditions found in HOT-CAT screens was validated by applying three promising catalyst systems in preparative hydration runs with each an aliphatic (1) and an aryl-substituted (4-tert-butylphenylacetylene; 18) acetylene as substrate. The reaction conditions were as in the screening, except for working at higher concentration. The system [IrCl(COD)] 2 À CSA (cf.

Entry
Catalyst (mol %) [b] Solv. [  Markovnikov regioselectivity of the iridium catalysts. The preparative run with 1 closely reproduced the screening result, and the yield of chromatographically purified 2 was close to the analytical value. Aryl acetylene 18 was hydrated with low chemoselectivity, and the yield was mediocre (Table 9, entry 2). Alkyne polymerization is the major side-reaction. Another catalyst system is Co(OAc) 2 À TPPÀ CSA in air, which is easy to set up and had shown favorable results in the screen (Table 4, entry 8). This was confirmed by the high isolated yields of 2 and 19, both of which approached the analytical yields (Table 9, entries 3,4). The third new catalyst tested is the simple Cu 2 OÀ CSA system. It performed equally effective in the hydration of 1 and 18, giving methyl ketones 2 and 19 in high yield. Chromatographic purification of the latter was complicated by the presence of methyl-acetal, thus 19 was isolated as crystalline dinitrophenylhydrazone (20).

Newly Detected Catalyst Systems from Screening Results
All of the established Markovnikov-and anti-Markovnikovselective alkyne hydration catalysts tested in the HOT-CAT screen returned a positive result (Table 1; Table 2 for Ru; Table 5 for Pt; Table 6 for Au). This provides a benchmark by proving that the screen positively identifies active catalysts. The high reaction temperature and short reaction time did not prevent observation of catalytic activity with heat-sensitive systems (e. g., Ag(I), Au(I/III)) whose thermal deactivation is reflected by lower conversions. The screens revealed many new findings, which can be divided into sub-categories based on activity and novelty, if compared to previously established catalyst systems: 1) Highly active catalysts with close analogy to established systems: High activity was detected for CoÀ TPPÀ CSA (90 % K; highly chemoselective; Table 4, entry 8; also cf. Table 9, entries 3 and 4), [IrCl(COD)] 2 (44 % K; low chemoselectivity; Table 8, entry 1) and Au(tBuPyPPh 2 )OTf (92 % K; Table 6, entry 7). Those are more or less obvious variations of reported catalyst systems (Co, [10] Ir, [56] Au [34] ). Yet there are novel aspects, such as the use of the simple TPP ligand with cobalt, as opposed to its watersoluble sulfonated version. The relatively high catalytic activity of [IrCl(COD)] 2 with acid (e. g., Table 8, entry 18), or of the complex alone (Table 8, entry 1), had not specifically been reported, although Ishii had devised the optimized system [IrCl (COD)] 2 À P(OiPr) 3 À ZrCl 4 in MeOH as highly selective alkyne hydration catalyst. [56] Their screening data infer that both P (OiPr) 3 and ZrCl 4 must be present to achieve high chemoselectivity.
2) Notable activity, moderate analogy to known systems: The activities of CuOTf·(PhH) 0.5 (Table S1, 160; 46 % K) or [Cu (MeCN) 4 ]PF 6 ( Table 6, entry 2; 16 % K) from the screen were regarded as very promising, since simple cationic Cu(I) had not previously been reported for alkyne hydration. A focus screen identified the co-catalytic effect of Cu(I) and H + in the practical Cu 2 OÀ CSA catalyst (Table 7; 85 % K). The system is active in aqueous methanol at low acidity and is thus different from recently published Cu(OTf) 2 -based systems, which work in acidic media with a preference for aryl-alkynes. [8c,69] An earlier Cu(II)system of Meier and Marsella in alcoholic solution [60] may well have contained a similar active component formed by reaction of Cu(II) with alcohol. [71] The complex cations [(P-P)Pt(C 6 F 5 ) + ] (18.5 % K at 37 % conversion with BINAP, 12 % K at 16 % conversion with dppe; Table 5, entries 4, 5) are chemically distinct from older platinum-systems based on PtCl 2 alone, or chloro complexes with alkene and phosphane steering ligands. [6,73] Those fluoroarylplatinum(II) cations were studied for alkene epoxidation catalysis by Strukul, [65,74] but now also appear as development candidates for catalytic alkyne hydration.
3) Low activity, at various degrees of analogy to existing systems: Significant alkyne hydration activity, although not at useful levels because of low chemoselectivity and turnover, was detected with (NH 4 ) 2 [OsCl 6 ] (13 % K, 67 % conversion; Table 3, Table 9. Preparative hydration with three selected new catalyst systems. [a] Entry Catalyst system [b] [ entry 15) as novel type of central metal in catalytic alkyne hydration (for stoichiometric precedence, see ref. [75] ). Transfer hydrogenation, isomerization and polymerization were major side-reactions. Further development is necessary to test if ligand effects can steer this activity further towards hydration. A glimpse of anti-Markovnikov hydration activity was seen with [RhCl(COD)] 2 (5.8 % A, 1.4 % K; Table 4, entry 14), but only in acetone, and at low activity and chemoselectivity. anti-Markovnikov selectivity was also observed for Cp*RuCl 2 À ISIPHOS in the 1-heptyne screen (Table S1, entries 68,  69), where the active species may have been related to known catalyst [CpRu(ISIPHOS) 2 ] + . [30,44] Albeit low, it is the first demonstration of such activity for a Cp*Ru system.

Transfer-hydrogenation, Alkene Isomerization
A family of metal-hydride induced reactions converts alkyne 1 to either terminal alkene 8 a and/or alkane 7 a by transferhydrogenation, presumably involving solvent as hydride-donor. Once the alkene stage is reached, double-bond isomerization to internal alkenes 9 may occur. Alkyne-semihydrogenation/alkene isomerization activity was distinctive with (NH 4 ) 2 OsCl 6 , where such pathways (including alkene hydrogenation) accounted for a remarkable 75 % in the solvent iPrOH (Table 3, entry 12). [79] Complex [IrCl(COD)] 2 apparently favors alkene hydrogenation of 1 to 7 a (Table 4), but since this was found alike in iPrOH (entry 19) and acetone (entry 20), an alternative explanation that 7 a was formed through a chain-walking redox-isomerization involving the -CH 2 OH unit may be considered. [80]

Reaction Development Experience Gained in the HOT-CAT Screen
The HOT-CAT screen keeps many reaction variables at fixed parameters and potentially deviates strongly from optimal or even satisfactory conditions. The development history of the practically applied catalyst systems (Table 9) provides critical case-studies: A first screen with cobalt was carried out in 2011, [81] prior to the report on water-soluble Co-porphyrincatalyst systems by Naka. [10] Seven runs (CoCl 2 , Co(OAc) 2 , vitamin B 12 ; each alone and with ISIPHOS; Co(OAc) 2 additionally with ISIPHOSÀ CSA; (Table S1, entries 101-107) displayed � 10 % conversion and no hydration product. The aim of the study at the time was to explore new alkyne hydration activity by harnessing potential accelerating effects of ambifunctional ligands, and the outcome for cobalt was negative. In this instance, the HOT-CAT screen failed to uncover a new type of activity, because a specific ligand (porphyrin type) was not considered for testing. More recently, and aware of Naka's results, we performed another screen (Table 4, entries 1-8) with cobalt. Vitamin B 12 displayed no activity with co-catalytic acid or iron(III)triflate. The switch to Co(OAc) 2 À TPP proved partially successful (Table 4, entry 6). The positive effects of oxygen and co-catalytic acid were then recognized in two additional experiments, providing the final catalyst system (Table 9).
With iridium, considerable activity was found with the simple metal precursor complex [IrCl(COD)] 2 in iPrOHÀ H 2 O ( Table 4, entry 19). Solvents were then varied (entries 20-22), and MeOH identified as best choice. The strategy to improve the low chemoselectivity through steering ligand effects in a focus screen failed (Table 8). Oligo-and polymerization were key distractors of this catalysis. The best conditions combine [IrCl(COD)] 2 with Brønsted acid, and this system was applied in two preparative runs, which still suffered from limited chemoselectivity ( Table 9).
The catalytic activity of cationic copper(I) was obvious in the initial screen with CuOTf (46 % K; Table S1, entry 160). Lowering the catalyst loading and temperature with [Cu(MeCN) 4 ]PF 6 reduced the yield to 3 %, whereas a solvent change to methanol increased it once again to 16 %. In spite of the low yield, the chemoselectivity was high, which was taken as a hint to "intensify" the reaction conditions (e. g., increase reaction temperature, -time or acidity level). A Cu(I) focus screen with Cu 2 O is a simple precursor for cationic Cu(I) with acids was performed (Cu 2 O + 2 HX = 2 Cu + + 2 X À + H 2 O; Table 7). The standard Brønsted acid additive of our screen, CSA [37] proved suitable, and, revealed a co-catalytic effect when used in excess (entries 4-8). The final catalyst system was suitable for preparative applications (Table 9).

Conclusions
We have proposed the HOT-CAT (homogeneous thermal catalysis; i. e. homogeneous catalysis at unusually high temperatures) screening approach as a tool to discover new catalysts systems for a predefined organic transformation. A peculiarity of this approach is that all test runs are performed at the same set of reaction parameters at a deliberately high temperature for a short reaction time. By fixing many reaction variables to constant values (temperature, time, concentration), and others to only few predefined settings (solvent type A-D, acid additive (yes/no)), the number of experiments in the screening of a variety of chemically distinct catalyst systems remains small. Concurrently, the short reaction time and standardized workup/ analysis render the procedure efficient. The bottlenecks are sample preparation and -analysis which require a skilled operator.

Utility and Limitations of the HOT-CAT Screening Approach
The HOT-CAT approach has proven suitable for broad-band screening across a large part of the periodic table. In spite of the elevated reaction temperature, there is neither extensive background reaction (e. g. by Brønsted acid) nor loss of starting material by non-productive decomposition. The screen is specific for catalytic activity by metal complexes and provides a wealth of information about the target reaction, accompanying side-reactions and about the catalyst systems tested. The number of side-reactions and side-products which were identified and analyzed for all individual catalyst systems in this study exceeds that which has been presented in published focused optimization studies for singular catalyst systems. Such data ensures comparability between diverse catalyst systems and will provide useful as reference in future searches for alkyne hydration catalysts.
At least two new promising catalyst systems of relevance for preparative application have emerged from the current work, namely Cu 2 OÀ CSA and Co(OAc) 2 À TPPÀ CSA in MeOHÀ H 2 O (4 : 1), which both were run under the same microwave conditions (160°C, 15 min) of the screening and provided 72-95 % isolated yield of product in 4 runs with an aliphatic and an aryl acetylene.
One of the motivations behind the screen was the search for new (non CpRu + ) types of anti-Markovnikov hydration catalysts. Notwithstanding indications of (low) anti-Markvonikov hydration activity with [RhCl(COD)] 2 (Table 4, entry 14), this search has not yet proven successful. The approach to invert regioselectivity by systematically combining ambifunctional pyridylphosphane ligands with metal precursor complexes was so far not effective.
Nevertheless, the HOT-CAT screen has allowed us to realize an unprecedentedly broad comparative overview of catalytic alkyne hydration and its accompanying reactions. We propose that as analytical tools progress, such generalized screens may be planned and performed as multi-reaction screens, in which the focus is no longer on a single target reaction, but as many reactions as analytical tools allow (cf. the concept of an undirected screen, ref. [5] ) In that sense, rather than screening for specific new catalyst systems, one will screen for reaction systems and optimize established reactions while searching for new ones.

Experimental Section
Unless otherwise specified, all reagents and solvents were obtained from commercial suppliers and used without further purification. Solvents used in catalytic reactions were purified by passing through a column of Al 2 O 3 and kept under an argon atmosphere. Column chromatography (CC) was performed on silica gel 60 (35-70 μm particle size), usually as a flash chromatography with 0.2 bar positive air pressure. Thin layer chromatography was performed on glass plates coated with silica gel 60 F 254 and visualized with UV light (254 nm) or Mostain (molybdenum stain with Ce(SO 4 ) 2 catalyst in H 2 SO 4 aq). NMR spectra were recorded at ambient temperature (19-24°C). 1 H NMR spectra were internally referenced to tetramethylsilane (TMS, δ 0.00) or residual solvent peaks (CDCl 3 δ 7.26; (D 6 )-DMSO δ 2.50). Measurements for qNMR analysis used 16 scans and a pulse relaxation delay d1 of 20 seconds. 13 C NMR spectra were referenced to TMS (δ 0.00) or solvent peaks (CDCl 3 δ 77.16; (D 6 )-DMSO δ 39.52).

General Microwave/qNMR Screening Procedure with Alkyne 1
Sample preparation: To a microwave glass vial containing a magnetic stirring bar, the metal precursor or -complex, the ligand, and/or any other additive were added. The glass vial was placed into a Schlenk vessel (29 mm diameter opening) and placed under argon by evacuating and filling with argon thrice. Solvent (2.0 mL), water (0.5 mL; 0.6 mL in case of solvent combination A with iPrOH) and 10-undecyn-1-ol (1; 50 μL, 0.26 mmol) were added to the vial in an argon counter-stream. The glass vial was closed with a corresponding standard cap and placed into the microwave reactor, where it was heated to 160°C with a holding time of 15 minutes. The cooled reaction solution was transferred into a 10 mL headspace glass vial (for GC analysis), which contained the internal standard for qNMR analysis (e. g., 20-25 mg of 1,3-dinitrobenzene). Ethyl ether (3 × 2 mL) was used to complete the transfer. The content of the capped headspace vial was homogenized by intense shaking. A quantity of 0.5 mL of the resulting solution was removed into a 5 mL headspace vial, where it was diluted with ethyl ether (2 mL) and washed with water (2 × 2 mL) [5 drops of sat. aq. NaCl were added to speed up phase separation] and sat. aq. NaCl (2 × 2 mL) by intense shaking and removal of the lower water phase by a Pasteur pipette after phase separation. The organic phase was transferred into a 5 mL headspace vial (with stirring) containing MgSO 4 (300 mg) and a magnetic stirring bar. The vial was closed with a butyl rubber septum and sealed with an aluminum cap. The suspension was magnetically stirred (1200 rpm) while vacuum was applied by placing a hypodermic needle connected to a water aspiration pump into the septum. A low vacuum (ca. 500 mbar, weak water flow in the pump) was initially applied to remove the majority of Et 2 O, after which the solid contents of the vial showed a dry, powdery appearance. A stronger vacuum (full water flow, ca. 20 mbar) was then applied for another 2 minutes. The solid residue was extracted with CDCl 3 (0.6-0.8 mL) by magnetically stirring for 5 minutes, and the suspension was filtered through a cotton plug into an NMR tube for analysis. Data analysis. 1 H NMR spectra were recorded with 16 scans using a relaxation delay (d1) of 20 seconds. After phase-and baselinecorrection, the relevant signal integrals as detailed in Figure S1 and Table S3 were collected for evaluation.