Org Prep Daily

September 20, 2009


Filed under: procedures — milkshake @ 5:38 pm


4-oxo-4,5,6,7-tetrahydroindole 5.15g (38.1 mmol; TCI-US) solution in anh THF 0.45L in a 1L flask was placed on ambient water bath and solid LAH 4.0g was added carefully, portion-wise, into the stirred solution, (4×0.5g then 2x1g – very exothermic, gas evolution!). The mixture was then placed on oil bath and refluxed under Ar on a 75-80C oil bath for 26 hours. The reaction mixture was cooled on ambient water bath, quenched by sequential addition of ethyl acetate 4mL followed by dropwise water addition, 4mL (very slowly – gas evolution), followed by 15% wt NaOH solution 12mL followed by additional water 4mL. The resulting slurry was stirred vigorously for 30 min, the salts were removed by filtration (the cake was washed thoroughly with THF) and the filtrates were concentrated on rotovap. The oily residue was distilled on highvac using a short-path distillation apparatus. (The material distilled pure; no fractionation was necessary, bp 57-60C/0.7 Torr).

Y=3.83g (84%th) of tetrahydroindole as an oily colorless liquid that gradually turns yellowish on light and air.[Note 1]

Anhydrous DMF 20mL in a 250mL flask was cooled on ice bath and neat POCl3 4.6mL (50mmol) was added dropwise (exothermic) under Ar. After 10 min, a solution of tetrahydroindole 3.83g (32.14mmol) in anh DMF 10mL was gradually added over 10min (exothermic) followed by additional anh DMF 2x5mL to wash the flask and the syringe. The cooling bath was replaced with ambient water bath and the reaction was stirred at RT under Ar for 13 hours (overnight). The reaction was quenched by addition of water 20mL followed by 15% wt NaOH solution 40mL. After 10 min, additional 15% NaOH 25mL was added. followed by water 100mL and the mixture was stirred vigorously for 30min on ambient water bath. The precipitated product was collected by filtration, compressed on the Buchner funnel, washed thoroughly with water, dried by suction and on highvac. Y= 4.282g (90.5%) of light tan shiny flakes

1H(d6-DMSO, 400MHz): 11.609(br s, 1H), 9.268(s, 1H), 6.679(s, 1H), 2.557(t, 5.9Hz, 2H), 2.446(t, 5.9Hz, 2H), 1.693(m, 4H)

Note 1: Tetrahydroindole is commercially available but is rather expensive. It has a strong unpleasant indole-like fecal odor – do not spill it outside the hood. Because of the oxidation-related darkening, it is best used within few days.


  1. This is not obvious to me, but why is the formylation specific to the 2 position instead of the 3 of the tetrahydroindole?

    Comment by T — September 20, 2009 @ 9:11 pm

    • The logical way to explain this is that for alkylpyroles the resonance canonics with multiple substituted double bonds give larger contributions. Therefore delocalization towards the “tetrahydro”-part predominates, hence 2 position substitution.

      As for indole, this would break the aromaticity of the benzene-part, and is therefore unfavorable, leaving only the delocalization to the “-position as major contributor.

      Comment by tora — September 21, 2009 @ 8:29 am

      • Being one of the lucky pyrrole chemists out there, this is obvious to me…
        If you write resonance structures for cationic attack (actually, nucleophilic attack by pyrrole itselve) you could end up with more delocalization, thus stabilizing. Also, part of the positive charge is located on the nitrogen (cfr imminium), and that also does the trick. There reactions do form small amounts of 3-isomer, but that tends to be less dan 2%.

        Comment by Jos Smos — October 23, 2009 @ 1:06 pm

  2. Its not obvious to me either but the general rule is that alkyl-substituted pyrroles get electrophilic ring substitutions into the 2 position whereas indoles go to the 3 position…

    Comment by milkshake — September 21, 2009 @ 12:28 am

  3. How was decarbonylation achieved with LAH?

    Comment by Rich — September 21, 2009 @ 1:41 am

  4. I think pyrrole benzylic alcoholates are so electron rich that they eliminate aluminum hydroxide equivalent by simple heating, and the produced stabilized benzylic cation, aza-fulvene or whatever this intemediate is, gets reduced one step further. In this way pyrrole carboxyesters can be taken down all the way to methylpyrroles. There is a similar reduction possible on EDG-substituted benzene rings as well, with Red-Al in refluxing xylenes

    Comment by milkshake — September 21, 2009 @ 1:53 am

    • U are right. I have used Red-Al to reduce the carbonyl gp, in the pyrrole (substtituted) ring to prepare a drug intermdiate.

      Comment by umesh — September 21, 2009 @ 2:48 am

      • a correction, I did reduce the side chain -CO- group using Red-Al. I agree with ur proposed mech…

        Comment by umesh — September 21, 2009 @ 11:24 pm

    • Any reason you didn’t use Et3SiH/TFA or similar? Good at complete reduction of aryl ketones, also by reduction of a benzylic cation formed from the intermediate benzylic alcohol. And, no refluxing LAH to burn down the house.

      Comment by mad scientist — September 24, 2009 @ 8:01 pm

  5. Any reason you didn’t reverse the addition in the first reaction, i.e. slowly adding the 4-oxo-4,5,6,7-tetrahydroindole to a suspension of LAH in THF?

    Just curious, insight appreciated.

    Comment by HB — September 25, 2009 @ 3:01 am

  6. I think alkyl pyrroles without any EWG groups on pyrrole would go to pieces, with the TFA/TES-H system. I used LAH because there was a good literature precedent for it.

    HB: sorry that Akismet spam guard ate your comment – I think it must have been the “lemonparty” in your e-mail address… There was no particular reason for the addition order, just the laziness. You are right, it would have been better if I dissolved the starting ketone in THF and then added this solution gradually into the excess of LAH – I saw some impurities in the crude product that could have been brought about by dimerization of the partially-reduced starting material

    Comment by milkshake — September 25, 2009 @ 8:17 am

  7. Alkylpyrroles are fairly basic and thus will be readily protonated by TFA (not on the nitrogen but on C-2). The resulting iminium can attack another pyrrole and make a mess, or, more likely, get reduced by silane, probably all the way to perhydroindole.

    LAH indeed works through azafulvenes, but prolonged reflux in THF is needed, which is always kinda scary. Making DMT is a way to start a good deal of academia lab fires. (Indole, oxalyl chloride, dimethylamine, LAH).

    Comment by Liquidcarbon — October 4, 2009 @ 5:38 pm

  8. Hi all, I have a problem which I was wondering if anyone was able to help with…
    I’m currently trying to synthesise 5-oxooctanoic acid by ring opening glutaring anhydride with n-prMgCl and catalytic CuI. The reaction is performed under dry argon, in THF with one equivalent of Grignard added. In addition to the desired product, I am getting the disubstituted tertiary alcohol, elimination product of this alcohol and also bizarrely the propyl ester of 5-oxooctanoic acid. The addition of grignard to anhydride was carried out dropwise and the reaction has been repeated at -40, -20 and room temperature- all with the same product ratios. Workup is addition of water, then 2M HCl with immediate extraction into DCM. All very confusing, especially as only 1 eq of grignard is used… any ideas? my only current thought is that the workup is always carried out at room temperature so any unreacted organometallic adds preferentially to the product rather than the anhydride at higher temperature. cheers πŸ™‚

    Comment by Artemis — October 29, 2009 @ 7:35 am

  9. This sounds about right, you are getting the second addition to keto because you have too little Cu catalyst. The elimination product happens from tert alcohol during the HCl workup. The propyl ester comes probably from unreacted propyl iodide in your Grignard – use an excess of Mg metal when you are making Grignard the next time. Definitely use more copper catalyst in the anhydride opening.

    You can even work with stoechiometric cuprate reagent – preferably generated by addition of 0.5-1 eq of CuCN to Grignard, a cuprate should leave the keto group alone, it is Grignard that does the over-reaction in this case. Cuprates generated from CuCN are more reactive than cuprates from CuI. (With cuprates of the R2Cu(-) anion structure the first alkyl transfers much faster than the second one. CN serves as a dummy ligand). Also if you have access to Scifinder, pull some old articles from Lipshutz on cuprate chemistry from about 20-30 years ago, he is the cuprate chemistry guru.

    I think the best would be actually to transmetallate Grinard into propyl zinc reagent in situ, by adding 0.55 equiv of anhydrous ZnCl2 (or ZnBr2) to the Grignard first, then do the anhydride ring opening in the presence of catalytic amount of Cu salt. (CdCl2 should work even better for transmetallation than zinc but Cd is somewhat toxic and you would probably have to make your own anhydrous CdCl2 by drying the hydrate with SOCl2). Please note that Zn reagents are more selective but less reactive than Grignards so you would probably need to use it between 0C and RT. Zn reagents are also fairly air-sensitive, unlike Grignards.

    The other alternative would be to first open the glutaric anhydride with MeNHOMe to make Weinreb amide, then do the reaction with grignard. Sometimes people use morpholine amides for the purpose but Weinreb amides work better.

    Comment by milkshake — October 29, 2009 @ 9:17 am

  10. thanks so much, i’ll try these out tomorrow πŸ™‚

    Comment by Artemis — October 29, 2009 @ 5:29 pm

    • Re: overaddition to the anhydride.

      There was a paper about a year back where they performed the addition of grignards to esters in the presence of a stoich. amt. of Et3N. They cleanly obtained the ketone with minimal tert. alc. formation. Their claim was that enolization of the ketone prevented further addition. Anyone ever try this out? Sounds a lot easier that transmetallating.

      Comment by Base Man — October 29, 2009 @ 8:13 pm

      • Thanks, I’ll give that a go also, seems like a nice method- im just looking for the ref now..

        Comment by Artemis — October 30, 2009 @ 7:25 am

      • I don’t remember what the recent reference is but I think there might be an older original ref (Synthesis 1980 pg 877) with the same procedure. Can’t access it now so I’m just going based on a snippet of an abstract. Anyway if you try the Et3N procedure let us know the results — I would be very interested to hear how it worked.

        Comment by Base Man — October 31, 2009 @ 12:30 am

  11. Also O(CH2CH2NMe2)2 works and how about this crazy way – you add 1 equiv. of MeON(MgCl)Me (or morpholineMgCl) to anhydride, then + 1 eguiv of your Grignard – in my understanding should work really clean without any byproducts – all one-pot. I am not big fan of Cu especially when you need 1 equiv.

    Comment by krest17 — November 1, 2009 @ 12:04 am

    • not a Cu fan? Your boss may find the lack of faith disturbing…

      Comment by milkshake — November 1, 2009 @ 2:11 am

  12. Well, as you probably saw – I am fan of Zn/Pd in water now :-). BTW, my boss is not as interested in Cu as before also.

    Comment by krest17 — November 1, 2009 @ 11:00 am

    • It makes no sense that organozincs survive – allyltrichlortin and alkylindiums in water in situ, why not – but Zn? It looks a lot like Mg to me. (By the way, I noticed your wife’s picture in the group photos – not bad at all).

      Comment by milkshake — November 1, 2009 @ 12:25 pm

      • Thanks man, not bad hmm… – beautiful I would say. BTW Boss was surprised as well :-).

        Comment by krest17 — November 1, 2009 @ 1:50 pm

      • I would write “excellent – her parents ought to be commended ” but I did not want to look like a stalker freak.

        So you brought the zinc charm with you, an influence from your old boss?

        Comment by milkshake — November 1, 2009 @ 2:14 pm

  13. Hi Krest17, Im guessing you’re alumni from Knochels group… Im quite keen to try the MeON(Me)MgCl method tomorrow morning so need to do some prep work- can the compound be made simply by reacting MeON with iPrMgCl.LiCl?? Cheers

    Comment by Artemis — November 2, 2009 @ 9:11 am

    • MeONHMe rather… :s

      Comment by Artemis — November 2, 2009 @ 11:23 am

      • morpholine should work as well

        Comment by krest17 — November 2, 2009 @ 11:26 am

    • Yep, that’s exactly the way – 1.05 equiv. of grignard or turbo-grignard. Would be reasonable to titrate it before.

      Comment by krest17 — November 2, 2009 @ 11:28 am

      • Thanks very much, the one-pot reaction worked a treat
        The second Grignard addition is a lot more sluggish than I would have expected though, room temperature addition of 2M n-PrMgCl solution took several hours to go to completion and a strong acid quench to kill the weinreb amide. No by-products, clean MS πŸ™‚

        Comment by Artemis — November 4, 2009 @ 4:50 pm

      • oh and the cuprate formed with CuI didnt work oddly enough, it gave a 2:1 ratio of mono to di addition. I might pursue it using CuI/LiCl or CuCN but as the Weinreb method worked so nicely I might just crack on with my synth- I dont think my supervisor was expecting this to take 4 weeks to resolve!

        Comment by Artemis — November 4, 2009 @ 4:52 pm

  14. Great! Always happy to help. (BTW, you was right about Knochel :-)).

    Comment by krest17 — November 4, 2009 @ 5:14 pm

  15. Hi MS,

    I am struggling to reduce a cyclic ketone into cyclic alkane. The system consists of cyclic ketone fused with aryl group (contains only fluoro group). We tried Wolff-kishner conditions for this transformation. But this reaction failed. Please suggest me any conditions for this reduction.

    Thanks very much


    Comment by marto — January 7, 2010 @ 6:32 am

    • If your molecule is not too sensitive you can try TFA+triethylsilane+BF3.etherate:

      The ketone reduction to alcohol is usually fast but alcohol tends to linger around; you need to monitor the reaction to ensure that the conversion of the alcohol to hydrocarbon is complete

      Then there is reduction of tosylhydazones that is preferable to classical Wolff-Kisner: TsNHNH2 in refluxing ethanol, about 1h, when complete conversion to tosylhydrazone cool down and add some NaBH3CN.

      Then there is Raney nickel reduction of cyclic dithioketal which should also work fine (HSCH2CH2SH and BF3.Et2O, then excess of Raney-Ni) but it is very smelly.

      Comment by milkshake — January 7, 2010 @ 4:57 pm

  16. Thanks very much MS.



    Comment by marto — January 7, 2010 @ 6:34 pm

  17. Hi MS,

    I am trying to reduce a hetero cyclic carboxylic ester to the corresponding alcohol derivative (it contains free NH also). This reaction is reported with LAH but I need to develop a scalable process with other reducing agents,

    I tried with Calcium borohydride but it does not work, please suggest me any other reducing agents or how can I mask the NH group during the reaction so that reaction can go smoothly.

    I need to develop a process which is economical and process relevance.

    Please suggest me if you have any experience



    Comment by marto — July 7, 2013 @ 10:39 am

    • I would try lithium borohydride in THF. If you need a cheaper reagent you can make it in situ with LiCl(vacuum oven dried) and NaBH4 in 1,2-dimethoxyethane

      Comment by milkshake — July 8, 2013 @ 10:02 am

RSS feed for comments on this post. TrackBack URI

Leave a Reply

Fill in your details below or click an icon to log in: Logo

You are commenting using your account. Log Out /  Change )

Twitter picture

You are commenting using your Twitter account. Log Out /  Change )

Facebook photo

You are commenting using your Facebook account. Log Out /  Change )

Connecting to %s

Blog at

%d bloggers like this: