With oil gushing into the Gulf of Mexico for almost three months, every attempt to stop the leak has failed, or fallen short - until now. Oil finally stopped gushing from the well on July 15. We look back at how we ended up here: what BP has tried and done so far.

July 20, 2010

Solution: Static Kill
Scientists are weighing a new option called  "static kill” for permanently sealing it. The "static kill" would involve pumping mud into the well to force oil back into the reservoir below. This is similar to the "top kill" method that failed earlier (see below), except that now the oil isn't flowing - hence the word "static."

Read more on static kill at Time.com

Engineers are proceeding with the relief wells that eventually will pump concrete into the well bore to kill it from the bottom. A static kill, if pursued, would hit it from the top.

Expectation
BP noted that the option could succeed where other similar attempts have failed because pressure in the well is lower than expected. Geologist Arthur Berman tells CNN's "American Morning" the relative simplicity of the static kill makes it an attractive option for BP.

Outcome
BP finished pouring cement down the well on Aug. 6, completing the job earlier than expected. The process took six hours. Retired Coast Guard Adm. Thad Allen said the cementing phase of the "static kill" operation is not the end of the process, "but it will virtually assure us there's no chance of oil leaking into the environment."

July 10, 2010

Solution: New better-fitting containment cap
BP said it was going to remove the old containment cap, replacing it with another that has a better fit. Robots removed six giant bolts from the apparatus July 11 so the new cap could be positioned.

Scientists will then be able to gauge the pressure inside the well and determine whether the cap is holding in the oil or if crews will need to continue siphoning oil.

BP says it will conduct a “well integrity test,” which involves closing the stack end and stemming the flow coming from the well.

If it works, oil collection via the vessels, Q4000 and Helix Producer, will cease. BP will then close in on the perforated pipe. This process, which will be done in collaboration with U.S. government officials, could take up to 48 hours.

Expectation
In the best-case scenario, the containment cap would have the ability to actually close down the valves and slowly contain all the oil – not plug the well.

If oil collection was still necessary, over the next two to three weeks, 60,000 to 80,000 barrels (2.52 million to 3.36 million gallons) a day could be collected as part of the containment process, BP Senior Vice President Kent Wells said. That's because the containment cap would allow four collection ships to access the well, rather than the maximum of three allowed by the old cap.

The oil giant said earlier as well that the cap "should improve containment efficiency during hurricane season by allowing shorter disconnect and reconnect times."

Coast Guard Adm. Thad Allen said he approved the cap-switch plan to take advantage of favorable weather predicted for coming days and because, once the switch is complete, the resulting capacity to contain oil "will be far greater than the capabilities we have achieved using current systems." Allen also stressed that once the capping device is on, "we would get the most accurate flow rate to date."

Outcome
The oil stopped gushing out on the afternoon of July 15 -  the first time BP has been able to gain control since the the Deepwater Horizon rig exploded three months ago and triggered the catastrophe.

The "well integrity test" began on July 15 after two days of delays, first as government scientists scrutinized testing procedures and then as BP replaced a leaking piece of equipment known as a choke line.

BP cautioned that the oil cutoff, while welcomed, isn't likely to go beyond the 48 hours. Valves are expected to open after that to resume siphoning oil to two ships on the surface, the Q4000 and Helix Producer, as government and BP officials assess the data and decide what to do next.

As of July 19, testing on a capped oil well in the Gulf of Mexico continues as the federal government says BP has addressed questions about a seep near the well. Retired Coast Guard Adm. Thad Allen, the federal government's response manager for the oil spill, says that a federal science team and BP representatives discussed the seep during a conference call, including the "possible observation of methane over the well."

See progress on relief wells

June 16, 2010

Solution: Second containment system
BP said Wednesday it has started collecting oil gushing into the Gulf through a second containment system attached to the ruptured well. The new system is connected directly to the blowout preventer and carries oil up to a second ship, the Q4000. The Q4000 uses a specialised clean-burning system to flare oil and gas captured by this second system. The Q4000 uses a specialised clean-burning system to flare oil and gas captured by this second system.

Expectation
This second system supplements the lower marine riser package (LMRP) cap containment system, which remains in operation, BP said. The new system is connected directly to the blowout preventer and carries oil and gas through a manifold and hoses to the Q4000 vessel on the surface.

Outcome
Oil and gas collected from the blowout preventer reached the Q4000 at approximately early on June 16. Operations continue to stabilise and optimise the performance of the second containment system.

See progress on the relief wells

June 3, 2010
Solution: An altered version of "cut and cap"
BP went back to the drawing board June 3 and planned to cut away the remains of the damaged riser pipe with a robotic-arm shearing device. A containment dome would then be put over the blowout preventer's lower marine riser package, but the larger, less precise shearing device will have left a rougher cut than what the diamond wire cutter was supposed to offer.

Expectation
The more primitive cut means that a rubber seal will not be as tight as previously hoped, so the dome may capture less of the oil. Nevertheless, the hope is that a good amount of the oil can be captured and brought to the surface until August, when BP is expected to be ready to use a relief well to seal the leaking well for good.Outcome
BP sliced off the remains of the damaged riser June 3, and Adm. Thad Allen, the federal government's response manager, called the news "extremely important." Robot submarines steered the new cap to the well later that evening. The cap placed over the top of the well funnels oil and gas to a surface ship, though oil is still spilling out from the cap and the valves.

On June 7, BP says that it has closed one of four vents on top of the cap, and that the process is working well. The company says it may not close all four of the valves because engineers think the valves may be releasing more gas than oil.

On June 10, scientists said as much as 40,000 barrels of crude are gushing into the sea every day. The previous estimate by researchers, made two weeks ago, was 12,000 to 19,000 barrels per day.

BP has collected about 73,300 barrels (about 3 million gallons) of oil since it placed a containment cap on its ruptured well, the company said.

June 1, 2010
Solution: "Cut and cap"
BP plans to send marine robots that will cut the "lower marine riser package,” or LMRP, on the well. This is a set of pipes that connect the oil well’s blowout preventer to the damaged pipe. After that, a diamond-cut saw will be used to make a "clean cut," preparing the way for a custom-made cap to be fitted over the package. One of a number of caps that BP has available, depending on the cut, will be placed over the package to bring the oil and gas to the surface.

The operation represents the first of three containment steps that BP plans to take. After the cap is on, a second operation will create a second flow through the blowout preventer, meaning there will be two channels of oil leaking to the surface.

Expectation
When asked about his level of confidence in the capping procedure, given the fact that other operations aimed at stopping the spill have failed, BP Managing Director Bob Dudley said the company has learned a lot from previous attempts. For instance, warm water will be pumped down in an effort to combat the formation of hydrates, or crystals, that blocked a previous containment vessel.

While the engineering has never been attempted at a depth of 5,000 feet, Dudley said, it is "more straightforward" than that used in previous operations. The cap "should be able to capture most of the oil," Doug Suttles, chief operating officer of BP's global exploration and production business said. However, he cautioned that the new cap will not provide a "tight mechanical seal."

If successful, the procedure will allow BP to collect most, but not all, of the oil spewing from the well. The cutting that precedes the cap placement carries with it a risk of increasing the oil flow, Dudley acknowledged. But "even with increased flow rate, this cap will be able to handle this." However, the BP statement said, "systems such as the LMRP containment cap have never been deployed at these depths and conditions, and their efficiency and ability to contain the oil and gas cannot be assured."

Outcome
BP's effort to slice off a damaged riser pipe stalled after the blade of a diamond wire saw got stuck June 2. The diamond wire cutter plan was dumped after the device got stuck midway through the pipe. It was freed and taken to the surface.

May 28, 2010
Solution: “Junk shot”
This method involves debris such as shredded tires, golf balls and similar objects being shot under extremely high pressure into the blowout preventer in an attempt to clog it and stop the leak. Engineers at BP used this technique along with the top kill.

What was expected
"Each of these [materials] has been proven to fill various-sized spaces in the blowout preventer until the flow is stopped," BP says in a statement on its website. "While there is no known perfect 'recipe,' a number of combinations of materials will be used." More drilling mud would follow the junk shot, with the hope that the two methods together would stop the oil long enough for cement to be poured into the well. BP’s Suttles compared the operation to stopping up a toilet.

Did it work?
This one failed, too. The process was carried out "a number of times" with the U.S. Coast Guard before the oil giant admitted that the experiment had failed, BP press officer Sheila Williams said. Engineers first used the junk shot to quell the 1991 Kuwait oil fires, but never at such depths. "I don't think we'll be using golf balls again," Williams said.

May 25, 2010
Solution: "Top kill"
The top kill involves pumping heavy drilling fluid into the head of the leaking well at the sea floor. The manufactured fluid, known as drilling mud, is normally used as a lubricant and counterweight in drilling operations. The hope is that the drilling mud will stop the flow of oil. Cement then would be pumped in to seal the well. The first round of pumping began May 26.

Expectation
Top kill has worked on above-ground oil wells in the Middle East but has never been tested 5,000 feet underwater. BP Chief Executive Officer Tony Hayward has given the maneuver a 60 percent to 70 percent chance of success.

A team of experts will examine conditions inside the five-story blowout preventer to determine how much pressure the injected mud will have to overcome. The company then performs diagnostic tests to determine whether the procedure can proceed.

Outcome
Three days of work involving three separate pumping efforts and 30,000 barrels of mud – along with what Hayward described as "16 different bridging material shots" – just didn't do the trick.

"We have not been able to stop the flow," a somber Suttles told reporters. " ... Repeated pumping, we don't believe, will achieve success, so we will move on to the next option." Suttles and other officials said that the top kill attempt to stop the flow did so – but only as long as they were pumping. When the pumping stopped, the oil resumed its escape.

May 14, 2010
Solution: Riser insertion tube
The riser insertion tube tool is a temporary solution that involves inserting a 4-inch-diameter tube into the Deepwater Horizon’s rise, a 21-inch diameter pipe, between the well and the broken end of the riser on the sea floor.

Expectation
The insertion tube connects to a new riser to allow hydrocarbons to flow up to the Transocean Discoverer Enterprise drill ship. The oil will be separated and shipped ashore.

Outcome
This seems to be the most successful effort thus far in containing some of the spill. The system was able to capture some of the leaking oil and pipe it aboard a drill ship, burning off some of the natural gas released in the process, according to a statement from the joint BP-Coast Guard command center leading the response to the oil spill.

The flow rate from the tube reached 3,000 barrels of crude (126,000 gallons) and 14 million cubic feet of gas a day as of May 20. BP’s Suttles said the company is "very pleased" with the performance of the tube. However, Gov. Bobby Jindal of Louisiana says the efforts haven't stopped oil from reaching his state's coastline.

May 12, 2010
Solution: Second containment dome or “top hat”
The "top-hat" cofferdam is a 5-foot-tall, 4-foot-diameter structure that weighs less than 2 tons and would be injected with alcohol to act as an antifreeze and keep its outlet clear.

Expectation
BP built the smaller dome after the containment vessel, designed to cap the larger of two leaks in the well, developed glitches. The new device would keep most of the water out at the beginning of the capping process and allow engineers to pump in methanol to keep the hydrates from forming, BP's Suttles said. Methanol is a simple alcohol that can be used as an antifreeze.

Outcome
BP abandoned the idea of using the “top hat” and opted to proceed with an insertion tube technique instead. It wasn’t clear why BP made that choice.

May 7, 2010
Solution: First containment dome
BP lowered a massive four-story containment vessel over the well to cap the larger of two leak points. The hope was that the container would collect the leaking oil, which would be sucked up to a drill ship on the surface.

Expectation
"If all goes according to plan, we should begin the process of processing the fluid and stop the spilling to the sea," Suttles said. But the method had not been done at such depths before.

Outcome
The plan was thwarted after ice-like hydrate crystals formed when gas combined with water to block the top of the dome and make it buoyant. The dome was moved off to the side of the wellhead and is resting on the sea floor, Suttles said. He declined to call it a failed operation but said, "What we attempted to do ... didn't work."

May 4, 2010
Solution: Drilling a relief well
The second well joins the failed well at the bottom, in rock 13,000 feet below the ocean. Once contact is made, drilling fluid and concrete will be put into the first well.

Expectation
This will lower the pressure on the failed well, enough to allow a concrete plug to be placed into it and permanently shut it down. The relief well could also be used for future oil and gas production. BP began drilling the second well this month, but it will take three months to complete. Weather conditions could prolong the process, Beaudo said.

Outcome
The well is expected to be completed by August. It would be a permanent solution to cap the leaking well.

End of April
Solution: Robots to shut blowout preventer
The rig’s blowout preventer, a 48-foot-tall, 450-ton apparatus that sits atop the well 5,000 feet underwater, failed to automatically cut off the oil flow after the April 20 explosion of the Deepwater Horizon rig. BP attempted to use remote-controlled submarines with robotic arms to reach access portals and activate the valve.

Expectation
The highly complex task was to take 24 to 36 hours, said Doug Suttles, chief operating officer of BP's global exploration and production business.

Outcome
Failed. "We've tried many different ways. Some things have showed promise; some haven't," BP spokesman Daren Beaudo said. "We don't know why the remote-operated shutdown systems haven't worked."