Home Glass Chemistry Jabo Glass Experiment

Jabo Glass Experiment

September 23, 2012

by Rich Shelby


I wanted to conduct an experiment to see if torching glass cane to make marbles would produce a similar “look” when the same glass used in torching was added to the big oven at a machine marble making company but in rod size glass rather then cane sized glass. I wanted to focus on reactive glass rods verses non-reactive rods. On September 21, 2012 I will perform this experiment at Jabo, Inc, located in Reno, Ohio. I want to thank the following individuals for assisting me in the experiment: Joanne Argabrite owner of Jabo, Inc., Dare Hawkins, Ronnie Ewers, Larry Burwell, John Parr, Curtis Hart, Casey Argabrite and Clint Argabrite.




Research Notes:

(1) COE is an abbreviation for coefficient of expansion. All solids have molecules that vibrate in place. COE of glass cannot be combined with other different COE numbers and if it is combined, usually incompatibility happens and the glass shatters.

(2) When one color cane vibrates at one speed and another color cane vibrates at another speed, there is a possibility of breakage when the glass cools if they are used in making a marble. There is a greater chance of breakage if one used a third, different color cane when making the marble. To avoid problems, the marble should be annealed. Annealing is when the mixed glass colors are allowed to soak at a high temperature for at least an hour allowing the molecules of the different glass colors to vibrate together. When marbles are annealed there is only a very small chance they will have a breakage problem.

(3) Usually, the annealing process is done in a kiln. Some people use a material called Vermiculite (a hydrous silicate of aluminum, magnesium, or iron) but this medium only allows very slow cooling and not soaking in very high heat necessary in annealing the marble. Some people think if one flame anneals the marble, using the flame of a torch, it is annealed but this is not true as flame annealing takes place for only as long as the person can keeps the marble in the flame of the torch and proper annealing must needs the marble to be in high heat for at least one hour or more.

(4) At marble making companies, they do not anneal using a kiln. Usually they place thousands of hot marbles in a metal barrel with a fiber blanket (fiber blankets allow for very slow cooling) to separate the marble layers. Because there are so many hot marbles in the barrel and the slow cooling takes place because of the fiber blankets, this allows for proper annealing - still the marbles on the outside perimeter may have breakage because they do not anneal properly but it is a great way to anneal large numbers of marbles in a cost effective way. Usually it takes overnight for this annealing process to occur.

(4) Reactive Glass:   Some glass cane contains metals that, when mixed together, produce different colors than the original colors of the cane. These canes are called reactive.

(5) Oven: An oven at a machine marble making company is a huge oven brick (bricks that can stand very high temperatures) “box” that is approximately 12’ across, 9’ high and 20’ long (the one I’ve seen at Jabo, Inc.). They come in different sizes but in this experiment I will use this size of oven.

(6) Cane is glass that comes in rods about the thickness of a pencil and come in various lengths from 13” to 42”. Rods that will be placed into the oven will have diameters from 2” up to 3” and are about 12” to 16” in length.

(7) Torching glass to make a marble has the possibility of getting up to 1450º F. The oven temperature may go as high as 2300º F. Torching uses a mixture of propane and oxygen to obtain the necessary temperature to melt glass canes. Because ovens get such a high temperature (about 2300 degrees F), the glass rods melt very quickly.

(8) A basic marble-making machine: The oven is heated to a very high temperature and base glass (the glass used before color is added to the oven and is usually white based) is placed into the back of the oven as it heats. The oven itself usually has three places where colored glass can be added: (a) at the back of the oven, (b) placed into holes on top of the oven called orifices or (c) at the front of the oven. At the front of the oven there are usually two streams of melted glass that flow out. Below these two flows is a cutter that moves left to right and back again and cuts the flow into cylinder shaped pieces that fall into two funnels. The cutter continually goes left to right then back again cutting the molten streams of glass. The very hot glass drops onto tapered rollers. The rollers are about 10’ long with groves in them where the hot glass moves and is rounded into spherical shapes. The marbles then drop off the rollers onto a flat, tilted metal surface where they roll downward hitting little baffles toward the big metal barrel where they will fall into to be annealed. It is quite impressive to see in action – all the movement and noise. I have only given a general description of the process.

(9) Researching the marble books, none of the following companies that used marble making machines (showing pictures of the marbles) used reactive glass as a general rule:

The Akro Agate Company, Alley Agate and the companies of Lawrence E. Alley, Alox Manufacturing Co., C. E. Bogard and Sons, Cairo Novelty Company, Champion Agate, Christensen Agate, Davis Marble Company, Heaton Agate Co., Jabo, Inc., Jackson Marble Company, Kokomo Opalescent Glass Company, Inc., Libbey Owens Ford, Marble King, Inc., Master Glass Company, Master Marble Company, M. F. Christensen and Son, Mid-Atlantic of West Virginia, Inc., The Peltier Glass Company, Playrite Marble and Novelty Company, Ravenswood Glass and Novelty Company, Vitro Agate Company and Vacor de Mexico.

(9) In the research study “Three-Dimensional Flow and Thermal Structures in Glass Melting Furnaces. Part I Effects of the Hest Flux Distribution Glass Science and Technology, Vol. 75, No. 2 pp 55-68.” from the Heat Transfer Laboratory School of Engineering Purdue University, I learned the study used a furnace very similar to Jabo, inc. oven. The flow and thermal structures in the molten glass bath of a typical glass melting furnace with a throat but without air bubbles or electric boosting is the same as Jabo, Inc. The major results of the study indicate (1) the heat flux distribution has no significant effect on the flow structure of the glass melt under the batch blanket where several Rayleigh-Benard cells develop in the span wise direction (see figure 1 and 1a below), (2) a heat flux gradient in the longitudinal direction is required to generate two recirculation loops in the direction (see figure 2 below), and (3) steep heat flux gradient in the refining part of the tank increase the refining recirculation loop near the front wall.

(1) figure 1 – a general oven diagram – compare this closely to Figure 2


FIGURE 1A – a general oven diagram - spanwise



(2) FIGURE 2 – a general oven diagram - compare this closely to Figure 1


(3) FIGURE 3 – a general oven diagram - longitudinally

(10)  the heat flux distribution shows no significant effect on the flow structure in the oven.

(11) Heat flux gradient, longitudinally, makes two recirculation loops. 

(12) Step heat flux increases the recirculation loop near the front wall where the stream is located.

(13) Jabo uses the submerged throat furnace.

(14) The roll cells tend to enhance the heat transfer from the fluid glass to the lower fluid temperatures.

(15) Sidewall heat losses is constant and heat losses at the bottom can either amplify or reduce the strength of the heat cells and make a viscous force within the glass melt.

(16) Tank depth has a significant impact on the arrangement of the heat cells under the batch.

(17) The melting of raw materials is a complex physicochemical process which involves a large number of chemical reactions and phase transformations occurring over a wide temperature range from 1000 to 1500 K.

(18) If one moves the heat flux toward the middle of the oven, the circulation loop drastically changes in size near the front of the oven.

(19) Moving the heat flux to the center will give more time for the refining and melting process to take place.

(20) Having the maximum heat flux neat the front, middle or back of the oven makes little or no affect on the streamline flow.

(21) Foam (dirt and bubbles) on the top layer of the oven acts to provide resistance to heat transfer as much as 60%.



Using a torch, If one uses reactive glass cane with non-reactive glass cane to make a marble will the results will be similar to the colors produced on a marble made with a marble making machine using reactive glass rods with non-reactive glass rods?


(1) Materials:

Prior to the run: Torch, propane, oxygen, marble making tools, kiln, and cane (both reactive and non-reactive).

At the run: use the marble making machine at Jabo, Inc.


(2) STEPS:

Prior to the run:

(1) Make 33 sample marbles totally using different combinations of reactive glass with non-reactive glass.

(2) Make 1 sample using only reactive glass.

(3) Make 1 sample using only non-reactive glass.

(4) Properly anneal all samples.

(5) Keep detailed notes.


NOTE: I just discovered that Jabo used a different white base glass than the base glass I made for the 35 sample marbles. They order from a different company that makes the white base glass they use.

Now one of two things will happen: (1) the type of white base glass matters and these 35 sample marbles are worthless to use testing my hypothesis. (2) the type of white base glass doesn’t matter and the reactive glass mixed with the non-reactive glass will give the same reaction colors regardless of the kind of white base glass used in the marble making process.

At the run:

My plan is to make 10 new sample marbles using Jabo’s white base glass before the actual run at the plant to determine the effect in testing my hypothesis using the same glass as was chosen for the run.

If the reactions are the same when compared with my samples I made before the experiment can go on. If the results are different, I’ll use the sample marbles I made the day before the actual run in my experiment and discard the 35 pre made samples I brought with me.

(1) Jabo and I will determine which 10 reactions are visually pleasing from the sample marbles and the cane (rods) of those colors will be added to the base glass in the oven during the run. These specific 10 color choices and their order will not be revealed to the public and the order of induction of said color rods into the oven and the exact ingress location will also not be revealed to the public. Understand the reason why this is to be as other companies may try to duplicate the results and copy the experiment.

(2) Each of the 10 reactions will be treated in 2 ways: (a) the first order of particular glass entry into the oven will be determined and kept the same with all of the 10 first trials (b) then the order of glass entry into the oven will be reversed for the second 10 reactions.

(3) Detailed notes will be taken when the above occurs.

At Reno Ohio:

On September 20, 2012, I was able to get the same base glass that will be used in the oven. I made 10 samples using the same glass that will be used in the oven as 10 new samples.

The results were very similar to the marbles I made before so the experiment will go forward.

In preparation for the run on Friday, an organizational chart was made. There will be 20 separate mini runs called 1A, 1B, 2A, 2B, 3A, 3B, 4A, 4B, 5A, 5B, 6A, 6B, 7A, 7B, 8A, 8B, 9A, 9B, 10A, and 10B. Whenever the mini run has an A label, it means the reactive glass rod will be placed into the oven first followed by the non-reactive glass rod. Then if the mini run has a B label, the non-reactive rod will be placed into the oven first and the reactive rod will be places next.

Unknown to me was cobalt blue powder was added to the oven about a week before the run. Too much powder was added and the base glass was contaminated so it stayed a blue color and not the white that was needed for the experiment to go forward. For about 10 days, running 24 hours a day, the oven’s glass stayed blue. On Saturday, September 22, 2012, the owner of Jabo, Inc. made an executive decision to drain the tank. When new white glass was added, it came out a slight shade of blue but was satisfactory to start the run on Sunday, September 23, 2012. Eventually all the blue was gone and the base glass stayed white.

At he first mini run, 1A, it was determined to drop the entire rod into the oven without breaking it into smaller shards. So that is what happened. In went the first rod into the oven and 3 minutes later, in went the second rod. In about 16 minutes color started to appear on the marbles and as time went on, more color appeared. But, the glass was too viscous and started to plug the stream.

A decision was made to break the rods up into smaller pieces, called shards and we did just that. Started mini run 1B with the smaller shards and later, the stream started to plug. So we decided to crush the rods even smaller even making powered glass. For the rest of the run, the crushed/powdered glass was used.

On September 23, 2012, there was another private run scheduled the same day. The Jabo Joker XXV Run. The experiment run started first: making 1A, 1B, 2A, 2B, 3A and 3B mini runs. Then the Jabo Joker XXV Run went next and when they were finished the second part of the experiment run took place.

Another factor became apparent when I did an inventory of the glass rods. I was missing 25 rods! The rods had gotten mixed-up with the Jabo Joker XXV Run’s glass so we got it all separated out but there was another problem. The glass company substituted glass to replace the glass they were out of with other glass rods we didn’t order. At this point my experiment failed. Having my control group not be what was intended made the experiment fall into the failed group. Nevertheless, making marbles would still go forward.

I was trying to see if the oven could match what I had made on the torch in 3 areas: 1) obtain several colors from using 2 different rods, one reactive and the other non-reactive. 2) obtain a color separation line that would outline the color band from the white base glass. 3) be able to control the width of the color band.

Of the 20 mini runs, only 2 held fast to my requirements of the experiment and that is not enough to make the experiment work.  

Now another thing happened. Because of how much base glass was used to get rid of the cobalt blue color, the amount of base glass that was used in the early morning experiment run, the amount the Jabo Joker XXV Run used, when we started our experimental run later in the evening, we ran out of white base glass. The owner started using a different company’s white base glass so we could finish the experiment.

Later in the second part of making marbles, we wanted to widen the color band and to have less white base glass show in the marble so we started to use more than 2 rods in the oven. Some of the glass brought to me had no identification and was crushed and added to the oven anyway. Accurate notes could not be taken other than just saying orange glass.

In spite of all the changes, marbles were made. The 3 things I wanted to achieve were successful. A good scientific experiment must have the ability to repeat is giving the same results. This one won’t but a lot was learned and I generally got what I predicted at the torch from the oven.

So, just for fun, here are 20 pictures of the marbles that were chosen by the owner for us to try to duplicate in the oven. When I made these before the run, I did ½ of the marble using reactive glass first, then non-reactive glass over the reactive glass. Then reversed the process for the other ½ of the marble. Then twisted the glass, laid it over itself and gave it a little more twist. In doing this, I can’t tell which side I did first or second in making the sample marble. Hence, at the run we have 1A, 1B, etc.

Sample marbles made before the Jabo, Inc. run:

Marble 1

Marble 2

Marble 3

Marble 4

Marble 5

Marble 6

Marble 7

Marble 8

Marble 9

Marble 10

Samples made at Jabo, Inc.

Here are some pictures of me making the sample marbles using the base glass that was used at the run:

              The Torch                                     Melting base glass                                                               Melting base glass


         Melting base glass                                             Heating the gather of base glass

                        Rounding in the mold                                                                             Checking things out

                                                                        Adding color                                                                                         Adding a different color


                                   Flame polishing                                                                                 Flame polishing


                           Close-up of a HOT marble - the colors will change as the marble gets to room temperature

Below are pictures of the 10 marbles I made September 21, 2012 using the same white base glass as Jabo used in the run. These were cooled using a fiber blanket.

Marble 1

Marble 2

Marble 3

Marble 4

Marble 5


Marble 6

Marble 7

Marble 8

Marble 9

Marble 10


(1) An error was made in not asking Jabo what type of white base glass they used. I assumed, incorrectly, that they bought the white base glass from the same glass making company that made the reactive and non-reactive glass cane and rods.

(2) I needed to have an oven expert involved prior to the run. My knowledge of what occurs in the huge oven when the glass is molten needs to be improved. I know about the physics of molten glass flow in a large volume but there is only so much a book can tell you. I need a person with years of experience in machine marble making and especially knowing what exactly happens in the 2300º F heat inside the oven to work with.

(3) Marble Sample numbers: 10 (not listed here) were chosen by the owner to make at the Jabo run.

(4) Rod colors that need to be ordered: (kind and amount not listed here) but were ordered from the appropriate glass making company.

(5) When the base glass ran out and another company’s glass was used, it seemed not to affect the colors in the color band.

(6) Having to rearrange the colors and use glass I hadn’t planned on using probably will give different results than predicted.

(7) This experiment will not follow the Scientific Method totally because of outside influence problems and solutions. These include the unavailability of certain glass rods, glass rods that were used to substitute those ordered and not having enough of the requested rod colors. Also, adding more colored rods later in the run with some not having any identification of the color or type. So, as a pure science experiment, it failed in its scope. That said, I was able to achieve all 3 things I predicted when I torched the sample marbles prior to the run so it was a personal success for me.

(8) The owner of Jabo, Inc. is calling the experiment marbles the Jabo Virgo Run because the run was made the last day of Virgo.

(9)  You can see the pictures of the final marbles under the Jabo section in the Jabo Virgo Run pictorial documentation.


I want to give a special thanks to Joanne Argabrite, owner of Jabo, Inc., for allowing me to use the marble making machine to see how reactive glass works with non-reactive glass.


Joanne also allowed me to do two other runs, separate from the experiment.

When I was 6, my mom poured this smelly clear liquid into a bowl then dropped in this tablet and it fizzed and turned the contents to a red color. Another bowl, another tablet and yellow appeared. Then another and blue appeared. She kept on making different colors in the bowls. Now I know she used vinegar and Easter egg dye. We made 24 Easter eggs using the colors in the bowls. She went outside to do something and in my 6-year-old mind I thought if I poured all the colors together in a big bowl I could make a rainbow! I am certain you know what happened when I poured all the colors into the big bowl. This ugly brown/gray color showed and I was SO disappointed I didn’t make a rainbow.


Joanne allowed me to put several colored rods into the oven to see what would happen. I was able to do this twice, once after the morning run and again after the evening run. Now my inner child is happy thanks to Joanne because I now have marbles instead of my rainbow.


Thanks again Joanne!!!!!