Friday, December 6, 2013

Dictation software: still not quite there yet for chemistry

Thanks to an excellent suggestion by Layne Morsch, I decided to put my iPad's voice recognition software to the test. Below is a recording of me reading the Evans aldol procedure to my iPad, and the untouched,  unedited text is below:

The boron out all reaction into the dry 2 L flask equipped with a large my next during bar is introduced 21.2 g 0.091 moles of the a slated oxazole the flask a sealed with a rubber septum and swept without you can solid is dissolved into hundred milliliters of anhydrous dichloromethane which is introduced via syringe. 
I thermometers inserted through the rubber septum and the contents of the Flasseur cool to 0°C with a nice bath to this cool solution is added via syringe 27 mL 0.107 moles of dibutyl boron true flight followed by 16.7 mL 0.1020 moles of triethylamine drop wise at such a rate as to keep the internal temperature below 3°C. 
The solution may turn slightly yellow or green during the dibutyl more Entre plate addiction and then to light yellow one triethylamine it's added the ice bath is then replaced with the dry ice chest that when the internal temperature drops below -65°C 10.3 mL 0.10 moles of freshly distilled benzaldehyde is added over a five minute. Via syringe the solution astir for 20 minutes in the drive sasktel bath then for one hour at icepack temperature. 
The reaction mixtures quenched by the addition of 100 mL of the pH 7 aqueous phosphate buffer and 300 mL of methanol to this Claudy solution is added by syringe 300 mL of a Toodle one methanol 30% aqueous hydrogen peroxide at such a rate as to keep the internal temperature below 10°C. 
After the solution Strid for additional one hour the vault materials removed on a rotary evaporator out of bath temperature of 25 to 30°C the resulting slurries extract with three 500 mL portions of diethylether the combined organic extracts are washed with 500 mL of 5% aqueous sodium by carbonate and 500 mL of brine dried over anhydrous magnesium sulfate filter and concentrated on a rotary evaporator to 435 to 36° 35 to 36 g White solid be on purified aldol adduct has a nice turmeric purity of greater than 97% as determined by gas chromatography the solidus recrystallize from circa 500 mL of 1 to 2 at the last eight hexane to yield 25.8 g or 84% yield of the desired aldol at with a melting point of 92 93°C 
the mother liquor is purified by flash chromatography column dimensions 8 x 20 cm with flash great silica gel.  The mother liquor is purified by flash on Aleutian with 25% of lasted hexing additional 2.8 g 9% die Stermara leaps dais there you Merickel he pure material is obtained.
I don't think Siri is quite ready yet to run a reaction in the lab, although perhaps I could use better diction.

Have a great weekend! Toodle!  


  1. Last I knew, even the fancy dictation software requires you to specify "comma" and "period" to get punctuation marks. That would help the readability somewhat.

    I'm fairly impressed that chemistry terms like "dichloromethane" are handled so well so often. You could have ended up with a more Anguish Languish ( transcript: "die Clara my thane" or some such.

  2. I would not even attempt this, simple spellcheck/autocorrect programs in text editors are bad enough. I had recently mistyped "flask" in a procedure that got auto-corrected as "flake", and since this was correctly spelled and in proper syntax, it escaped my attention (and spellcheck's) and got into the final document

  3. I think Siri's experimental may just work, but I'm not sure what a slated oxazole is.
    Anyway is it worse than you see in some of the publications these days?

  4. Looks eerily like the experimental procedures I've been getting from our Chinese CRO.

  5. George Whitesides uses Dragon to write grants and send emails to colleagues.

  6. Who is "Claudy" and why do they get their own solution named after them.