Sunday, November 27, 2011

How does a chapati or roti fluff up?

Many of us are familiar with the way to make a perfect chapati or roti:  (i) roll flat a piece of moist dough evenly, (ii) heat one side on a pan, (iii) flip and heat the other side, (iv) then put it directly onto the flames.  If done right, the chapati puffs up like a balloon, a giant bubble of water vapor contained within the intact surface of the roti.  Why and how does this happen? Centuries of chapati making have led, through trial and error, to this sequence of steps. Think through this, as we did through a dinner conversation with my younger son, and there are clear scientific reasons behind each of these steps.

When a chapati is heated, dissolved water in the dough evaporates and collects together to form small bubbles which then coalesce into one large bubble. As more dissolved moisture is converted to vapor, the bubble grows, pumping up the chapati and pushing the malleable dough outward.  Instead of growing large as it does, why doesn't the water vapor bubble simply escape by perforating the front or the back surface, as in the figure below?  These surfaces are readily available to the bubble less than a millimeter away!

The answer lies in the manner that the chapati is heated.  Take the hottest part of the chapati—the bottom surface that is in contact with the hot pan.   Water vapor that has formed within an escape depth of this surface leaks out through the back.  This water denuded outer layer of dough now hardens, losing its malleability and making it harder for a  water vapor bubble to punch through.
Now the top surface needs hardening as well—which is why in the best puffed chapatis you need to heat one side, then flip over to heat the other side.  Once both surfaces have been hardened, the escape routes from top and bottom have been sealed, the growing bubble is frustrated and forced to expand laterally, as in the figure below.

Why is there a single bubble and not many?  When a bubble is formed, a new surface is created (i.e. the circumference of the bubble).  This has an energy cost because atomic bonds have to be broken to create the surface (if you split a wooden log in two, you have expended energy to break atomic bonds to create two new surfaces that weren’t there).  Systems in nature always try to minimize the energy cost if they can. If two bubbles of the same radius coalesce to form a single bubble, the surface area reduces to 2-1/3 of the original area, if n bubbles coalesce, the area reduction is a factor of n-1/3.  Creating a single large bubble instead of many small ones therefore requires a smaller area of surface to be created, and a lower energy cost.  It is thus fated that the bubbles must give up their individuality.  This is why in a good chapatti, in the end, there is one huge bubble. 

We are still left with step (iv), that the best chapattis are made when—after heating each side on a pan--the chapatti is put directly onto the flames for a vigorous fluff up.  There is a reason behind this as well. The dough is not perfect.  Some water vapor can still leak out through the top and bottom surfaces.  There is therefore a small “leak rate”.  Now when the dissolved moisture is converted to vapor in a hot chapatti, there is a certain gas generation rate. If this generation rate is much larger than the leak rate, then we can expect a good puff, otherwise not. Sticking the chapatti directly onto the flames raises its temperature and rapidly converts the moisture into gas—greatly increasing the gas generation rate and making the leak rate negligible in comparison.

Thursday, November 17, 2011

Indian food, the Ngram viewer, and some statistics

I heard an interesting talk recently by Jean-Baptiste Michel from Harvard’s Cultural Observation Laboratory, on tracking the appearance of specific words in print since the early 1800s.  Their database was the vast collection of digitized books, that now cover about 4% of all printed books in English.  The results offer a window into cultural trends through the ages, and has the potential to become a useful quantitative tool in the hands of social scientists.  This tool is available on the web ( and for some amusement, given that this is an Indian food blog, lets take a quick look at how Indian food has been received in the West through the years.
Blue: tandoori; red: vindaloo; yellow: raita, green: mulligatawny

Take a look at the graph above.  The Y axis shows the frequency (as a function of percentage of all printed words) with which the words tandoori, vindaloo and raita appear in all English books printed since 1800 that are available via Google Books.  I picked these words as being representative of Indian dishes popular in the West.  They make their appearance in print around the 1960s, coincident with the wave of sub-continental immigration into the UK.  The curves rise rapidly, reflecting the popularity of Indian food the past 15 years.  Compare them against the granddaddy of Anglo-Indian cuisine—the Mulligatawny Soup.  Nobody has Mulligatawny soup these days, we never had it growing up in India, and it is largely unavailable except in some Indian restaurants in the US (and perhaps England) who strain to create a “Raj” ambience.  But at one time in the 1800s, Mulligatawny was king of the hill, and one of the first culinary products that came out of the British –Indian encounter.  Consider this advise given to a young man considering a commission with the East India company in “The Surgeon’s Daughter” (1800) by Sir Walter Scott,
 “'If you, my dear fellow,' continued he, extending his hand to Middlemas, 'would think of changing sheep-head broth and haggis for mulligatawny and curry, I can only say that, though it is indispensable that you should enter the serv-ice at first simply as a cadet, yet, by , you should live like a brother on the passage with me; and no soonef were we through the surf at Madras than I would put you in the way of acquiring both wealth and glory…
And, Walter Scott was not the only one--the graph shows a steady representation for Mulligatawny since the 1820s that has continued to this day.

How popular is Indian food compared with Chinese?  The figure below shows the frequency of Chow Mein, Chop Suey, and Lo Mein over time. Chinese food of course became popular at an earlier time compared to Indian--this is supported by the graphs; what is also interesting is the relative use of the words themselves. Chop Suey is an older term, was at the height of its popularity in the early 1940s for reasons that are unclear (perhaps related to WWII?). Its usage has dwindled, though it is still popular today, at about the same level as Chow MeinLo Mein camd later, around at the time that Indian food started becoming popular.  I would have thought that Chinese fare is still ahead, but we find that the popularity of tandoori and Chow Mein--at least in print-- are now about the same.

Blue: chow mein, red: chop suey; green: lo mein

And, digressing a bit from food, what about Indian stereotypes?  Even into the 70s and 80s, the “Indian snake charmer” kept alive the colonial concept of the mysterious East; it was even pandered to—I recall Mad Magazine from the 70s with a spread from the well known Indian cartoonist, Sudhir Dar, which had an entry that played on this same theme.  This embodiment, as we all know, gave way to the Indian programmer in the 2000s—even making their way into TV commercials the past few years.  So how do they look on Ngram? Surprisingly, the Indian snake charmer is still around, after enjoying some peak attention between 1900 and 1920 for reasons unknown.  The Indian programmer has climbed steeply since 1990, though unexpectedly, still neck and neck with the snake charmer .
Blue: Indian programmer; red: Indian snake charmer